[Six years ago, as a demonstration, I wrote two five-paragraph essays: "Composition" and "Decomposition". I wrote them for the exercise, to demonstrate something formal. But I just reread them and kind of liked them in themselves. Here's what happens when they're combined into one.]
Composition is the art of constructing texts. In his classic, if somewhat forgotten, little handbook, Rhetoric and English Composition, Herbert Grierson points out that this can be understood on three levels: the construction of sentences, the construction of paragraphs and the construction of whole texts. But he also emphasizes the relation between these levels. Not only is the "the ideal paragraph" essentially "an expanded sentence", the work should always be guided by the same principles. At all levels, "coherence and the right distribution of the emphasis as determined by the purpose you have in view" are paramount. There is a sense in which style is just your "choice of words". Composition demands that we put words together, in sentences, paragraphs, and texts, to achieve a well-defined goal.
In a sentence, words are put together grammatically in your attempt to mean something by them. In isolation, words don't mean anything very specific; they do not convey a clear meaning. In fact, until a group of letters is positioned among other words, it is unclear even what language it belongs to. The word "hat", for example, refers to something you wear on your head in English but is a form of the verb "to have" in German. A word really only finds its meaning in the context of a sentence, and here its meaning is determined by usage. Usage is the governing principle of grammatical correctness and that is why the way you construct your sentences goes such a long way towards defining your style. What is often called "accepted usage" by grammarians and editors determines the effect that particular words have in particular combinations and in particular settings. The style of your composition, as you try to get the words to mean what you want to say, is your struggle with what usage (in your particular context) would have your words mean before you started using them. This struggle takes place first and foremost within the sentences you write.
If a sentence is an arrangement of words, a paragraph is an arrangement of sentences. There is obviously no grammar of such arrangements, but there are some principles to keep in mind. First and foremost, a paragraph should have a unified purpose. This means that all the sentences that are gathered in a paragraph should, at a general level, be about the same thing. They will not, of course, say the same thing, but they will each play a specific role in elaborating, supporting or illustrating a common subject matter. This, in turn, is but one part of the overall subject matter of the text. "The bearing of each sentence upon what precedes," says Grierson, "should be explicit and unmistakable." In an important sense, then, the text's agenda is not advanced (moved forward) within its paragraphs but between them. A paragraph slows down and dwells, as it were, on a particular element of the larger subject covered by the text.
Ultimately, a composition consists of a series of paragraphs. If you looked only at the topic sentences (usually the first sentences) of these paragraphs, you should get a good sense of how the text is organized and what it wants to accomplish. When writing a text, it can therefore be useful to generate an outline simply by listing these topic sentences and perhaps to organize them further using what will turn out to be section headings. You will here need to decide what the organizing principle of the text as a whole will be: a narrative plot, a logical argument, a call to arms, a set of impressions, etc. "It is," says Grierson, "an additional satisfaction if in an essay or a book you can feel at the end not only that you have derived pleasure from this or that part of the work, or this or that special feature—the language, the character drawing, the thoughts, the descriptions—but that as you lay it down you have the impression of a single directing purpose throughout". The reader should feel, as Aristotle also said, that there was a reason to begin exactly where you began and end exactly where you ended. The composition of the whole text depends on the way the paragraphs are strung together to achieve this single purpose.
Texts are constructed out of words, not ideas, as Mallarmé might say. Words are arranged into sentences, sentences into paragraphs, and paragraphs into whole compositions. The correctness or rightness of these arrangements depend on their overall effect, that is, their aptness to a single purpose. This purpose, which gives the composition its coherence, makes demands of the text as a whole, and the demands of the text will make demands of the individual paragraphs, which will then pass further demands onto the sentences. It's really like any other construction project: the smaller parts must contribute to the larger whole; they must make themselves useful. It is often in working with the sentences that one discovers the style that is best suited to accomplishing the overall goal, always working under the general constraints of usage. It is also here that you might find a truly creative solution to the problem of writing, which can be a very complex problem because there are so many different reasons to write. Composition, in any case, is the simple art of solving it.
But is it really so simple? Grierson insists that good composition is characterized by "coherence and the right distribution of emphasis as determined by the purpose you have in view". But who are "you"? Grierson clearly assumes that the writer, operating somewhere well outside the text (somewhere beyond the page on which the words have been gathered) is in control of his (always his) expression. He would no doubt install the reader in the same space. But why, then, do these two subjects (of the same merciful lord) need a text? Couldn't "you" and "I" just talk to each other? Can't we all just get along? No, let us assume that the only "you" to speak of is the reader. Texts often crumble in our hands when we pick them up. If "composition" denotes how a text is "put together", "decomposition" might denote how they "come apart". If construction is about how a text is built up, how it is assembled out of words, sentences, and paragraphs, deconstruction is about how a text breaks down, how it collapses, as Derrida taught us. Decompsition is about activating the incoherence of the text, its excesses of emphasis, the indeterminacy of its always multiple points of view.
A text coheres if it is read charitably, that is, morally. Cued by markers that suggest the text wants to describe a place, or tell a story, or put forth an argument, we let our familiarity with space, time, or logic respectively, (and always respectfully) inform our reading. Herbert Grierson emphasizes the we have "knowledge by aquaintance" of these "orders of phenomena", that is, we are continually aware of these orders in going about our ordinary business. Coherence is an attribute of the surface of discourse. The first sign of the underlying incoherence of a text is therefore the superficial interference, or dissonance, that may be observed between spatial, temporal and conceptual orders. The story may at first seem plausible, but not in the place suggested. The arrangement of things in the room may be quite reasonable but how did they get there?
All sorts of embarrassing details lurk in the clash of orders that deconstruction brings to the fore. Most important, however, is the order that Grierson leaves out, or (more charitably) subordinates to the order of thought (logic): the order of emotion. Words and sentences do not just evoke thoughts, facts and acts, they also evoke particular feelings. Too often, writing makes too little or, in other literature, too much of the emotional response of the reader. It underestimates the indignation or overestimates the emphathy of the reader. And we, as readers, often much too easily play along. "[The] law of coherence is a heuristic rule," said Foucault in The Archaeology of Knowledge, "a procedural obligation, almost a moral constraint of research." It tells us
not to multiply contradictions uselessly; not to be taken in by small differences; not to give too much weight to changes, disavowals, returns to the past, and polemics; not to suppose that men's discourse is perpetually undermined from within by the contradiction of their desires, the influences that they have been subjected to, or the conditions in which they live.
To decompose a text is precisely to confront it, not with the "order of phenomena" normally supposed by the reader (to have been intended by the writer), but with the disorder by which the text is strangely disposed. It happens whenever we shamelessly insist on reading the text.
Deconstruction is a shift of emphasis while reading. It actively challenges the principle of composition: "coherence and the right distribution of the emphasis". We have just dealt with coherence; to better understand the decomposition of emphasis, consider two different ways of playing Bach. Wolfgang Sandner has said that Keith Jarrett plays Bach "emphasizing nothing, demanding nothing, concealing nothing and withholding nothing. In one word: natural." He cites the pianist himself in support of this thesis. "This music does not need my assistance," says Jarrett. "The melodic lines themselves are expressive to me." Compare this with what Sandner says of perhaps the most famous interpreter of Bach, Glenn Gould. "Obviously," writes Sandner, "he did not even trust his own analyses. He remained in search of clues. He spread the tones, loosened their coherence, emphasized side-lines and with his extreme tempi subjected the works of Bach to a kind of stress test."
There may be no better way to summarize the spirit of deconstruction: don't trust your own analyses but continue the search for clues; emphasize side-lines and read at extreme speeds (whether fast or slow); all in all, subject the text to a stress test. You can experience the difference by listening to their recordings of the thirteenth prelude in Book I of Bach’s Wohltemperierte Klavier. By slowing it down, and emphasizing the space between the tones, Gould is able to draw our attention to our own contribution to the music, our listening. It is important to keep in mind that Sandner is talking about two performances of the same composition, two "readings" of the same "text". The composer may have preferred one or the other, but there is no basic sense in which one is "right" and the other "wrong". Each reveals something about the composition. A "natural" emphasis may offer a great deal of immediate aesthetic pleasure, to be sure, but deconstruction is the pursuit of a more difficult beauty. Decomposition results from an excess of emphasis.
It often assumed that good academic writing is rooted in a singularity of purpose. "The specialist," Grierson tells us, "need think of nothing in regard to style but clearness and precision." And he alleges a reason: both his subject-matter and his audience is given to him so his point of view is largely fixed in advance. He need only ensure that his style does not obstruct the audience's view of his subject. "Everything else is an intrusion, and an unnecessary intrusion, because he can count upon willing and patient readers who desire to study the subject". For Grierson, specialist writing is a particular way of establishing the point of view of a text, which in turn "determines everything". Since, following Aristotle, the point of view depends on the speaker, the subject-matter, and the audience involved, says Grierson, there is really an infinity of possible points of view for any text.
But he makes a crucial assumption. A given text, he notes, will have a single point of view; the writer can make a series of rhetorical decisions to, as it were, "fix" it. Deconstruction draws this assumption radically into question, beginning with the allegedly singular purpose of the writer; for even the most academic writers are torn, at least, between enlightening their readers and furthering their careers. This immediately suggests multiple audiences, but it also suggests that a text is about any number of things that are not mentioned in the abstract. Deconstruction attempts to chronicle the "wars of signification" that take place behind the often irenic facade of an academic text. What we might call "academic composure" is fostered by an illusion of the writer's singular purpose, namely, that his only intention is to instruct a "willing and patient" reader, one whose only desire, in turn, is "to study the subject". Once we drop this assumption the text begins to decompose.
The essential thing is to read the text. To deconstruct it, we loosen its coherence, redistribute its emphasis, and question the unity of its purpose. All of these are acts of reading. It is true that deconstruction demands that we set aside the usual obligations of reading; it demands that we read against what are often the clearly marked intentions of the author. But deconstruction should not be taken as a personal attack on the author. Grierson assures the writer that the text will be read in the light of the reader's "knowledge by acquaintance" of the basic orderliness of experience, that it will be read with a natural emphasis, that its readers, desirous only of study, will be patient and willing. Such assurances, when believed, produce a particular kind of text, and it may be a very good one. Every once in a while, however, we need as writers to see what our assumptions about the reader have actually accomplished. On such readings, the text will begin to come apart, sometimes like a collapsing structure, and sometimes like a mound of compost. We can use the results of such decompositions when we compose texts of our own.