Parkinson’s Law states that work tends to expand to fill the time available for completing it. A variation might describe the elasticity of language: words can be multiplied to fill the space available for expressing an idea. In light of this law, most writing textbooks preach the virtue of conciseness, especially those that focus on technical writing (e.g. Burnett 309-12; Lannon 263-73; Reep 156-57; Shelton 17-22; VanAlstyne and Maddison 14-16). A typical section in such texts demonstrates how to recognize and eradicate problems like redundant pairs (“benefits and advantages”), redundant modifiers (“man­datory requirement”), redundant categories (“rectangular in shape”), phras­es used where words would do (“at this point in time” instead of “now”), and empty sentence openings: ‘‘There is a strong likelihood of rain tomorrow," rather than “Rain is very likely tomorrow.” Dietrich Rathjens identifies brevity as one of the definitive components of clarity in technical writing: “Brevity means that we cut every word, phrase, or sentence, whenever possible, to the extent that the intended meaning of the text is not altered” (43). And for well over a decade, Richard Lanham has been urging writers to calculate and reduce the “lard factor” in their writing (4).

Collectively, the sins listed above are called redundancy or wordiness. Redundancy is widely seen as a kind of linguistic cholesterol, clogging the arteries of our prose and impeding the efficient circulation of knowledge. However, I will argue that, just as a more thorough understanding of cholesterol reveals the existence of good cholesterol (HDL) as well as bad (LDL), so a broader view on the principle of redundancy reveals its effective­ness in certain situations, particularly beyond the sentence level. In this article I aim to revive the beneficial or functional sense of redundancy and show that functional redundancy in writing need not be a contradiction in terms. I believe a discussion of redundancy should include its opposite, ellipsis, so I will define both terms, emphasizing the beneficial sense of each, and then show how they appear in both reading and writing. In the latter part of the article, to illustrate the pervasiveness of redundancy and ellipsis, I will discuss examples of each in document design and in figures of speech. My attention will mainly be on technical writing, but the principles I will discuss may apply to other genres, too.

The gist of my argument is that, while less is often better, using as few words as necessary does not always mean using as few words as possible. Students should learn to eliminate needless syntactic redundancy in the context of a broader theory of how and why to create redundancy or ellipsis—a theory which considers how readers read, what they need at certain points in a text, how redundancy and ellipsis relate to each other, where they appear in language, and how they affect the reading experience. I offer the ideas discussed below as notes toward such a theory.

Redundancy and Ellipsis Defined

Redundancy exists when more of the same thing is available, possibly in another place or in another form. Redundancy involves reinforcement at a level above the required minimum. It is characterized by amplification, repetition, restatement, or the presence of comparable alternatives. Contin­gency plans create redundancy, as do spare keys and tape backup drives in computers. Technical writing students bound for careers in engineering—particularly structural, mechanical, or electrical—may already be familiar with redundancy as a sound design principle used to improve reliability. If the failure of a system is critical or likely, the parts that are most prone to failure will have functional redundancy—a margin of excess capability—built into or around them. Critical components will be duplicat­ed or designed so that they can be bypassed if they develop a problem. In this way, local failure of a part need not cause catastrophic failure of the whole system.

The drawback of redundancy is that it reduces efficiency, adding weight and expense to a design, but in a world where systems cannot be continuously monitored and will almost inevitably fail—in a world governed by Murphy’s Law as well as by Parkinson’s—redundancy remains a valuable method of ensuring reliability in a variety of engineering applications. If all oil tankers were built with the redundancy of a double hull, the “Exxon Valdez” might have escaped infamy. NASA’s Atlas II rocket has a Launcher Stabilization System that keeps it steady during the critical period between the removal of the mobile service tower and liftoff. Because hydraulic systems are relatively unreliable, and because so much is at stake, the system has two individual hydraulic circuits on each side of the launch vehicle. If one leaks, the other will maintain pressure (Nakamura 33-35).

I offer these examples to make the point that redundancy can be a vital safeguard, and to suggest that it plays a similarly important role improving the reliability of technical communication. The technical writer is seldom available for readers to consult if the readers are confused, and misunder­standings obviously may have disastrous consequences. Furthermore, as I will explain later, technical writing is often read in a way that increases the chances of misunderstanding.

While most textbooks urge conciseness, there have been a few scholars who have swum against the stream. Nevin Laib argues for teaching amplification:

The opposite of redundancy is ellipsis, defined broadly as the omission of material that observers might expect to encounter. They may be able to infer what is missing, or they may just be aware that something has been left out, in which case observers become aware of a corresponding ellipsis or gap in their knowledge. To illustrate with an anecdote, while visiting the ruins of Conway Castle on the north coast of Wales last summer, I noticed a pattern of holes on the inside surface of the stone walls. I saw these holes as ellipses, inferring that something had once filled them, but I could not guess what, which indicated an ellipsis in my schema for castle architecture. The guide was able to fill that ellipsis in my knowledge by explaining that the holes had been used for wooden scaffolding. Readers recognize ellipses in text when they notice that words or ideas have been implied or invoked but not stated, and they notice ellipses in their own knowledge when they are unable to fill the ellipses in the text. When I referred to Murphy’s Law three paragraphs ago, the sentence contained a highly predictable syntactic ellipsis (the omission of “Law” after “Parkinson’s”). Whether my reliance on readers to supply the definition of Murphy’s Law constituted an ellipsis of information is debatable. Because syntax follows widely held conventions familiar to most readers, sentence-level ellipsis is relatively easy to spot. On the other hand, ellipsis of information or meaning is a function of readers’ background knowledge of subject matter rather than language, and that knowledge can vary widely amongst readers. I will return to this point later in this discussion.

Redundancy and Ellipsis in Reading

A text becomes redundant when it represents or invokes information that readers already have, either because they knew the information before reading the text or because it was presented earlier in the text. This definition helps show, as Alice Horning has argued, that coherence and readability require a certain level of redundancy. The presence of overlapping informa­tion or cohesive ties between sentences creates redundancy, which contrib­utes to the coherence of a text by adhering to the given-new contract, presenting readers with new information in the context of familiar or given information (Horning, “Readable Writing” 137-141; Psycholinguistics 65). For a text to be readable, there must be some overlap of information within the text and some overlap between the text and readers’ prior knowledge. The more easily readers can relate new information to what they already know, the more easily they can understand and retain that information.

Frank Smith and other reading theorists have explained that readers make good use of the inherent redundancy of language—the existence of alternative cues that suggest the meaning of what is on the page. Redundancy lets us recognize a word on the basis not only of its individual letters but also our knowledge of its shape, sound, syntactic context, and semantic context (Frank Smith 56-57). Fahnestock explains how linguistic knowledge pro­vides the redundancy we need to understand unfamiliar technical discourse by “reading backwards” (“Connection”). So, even writing judged to be very concise is always functionally redundant in the sense that competent readers can find in it multiple sources of information to help them recognize the words and the meaning. Beginning readers have difficulty because they have not yet orchestrated all these multiple, redundant pathways to meaning. We can say that the redundancy is built into the language, but it is also true, as Smith points Out, that readers can only take advantage of that redundancy on the page if they have the relevant background knowledge:

I want to argue now that redundancy and ellipsis can also be found in the way readers process text. If there is enough redundancy in the text or in the reader’s knowledge, the reader can shift into a high gear and employ elliptical reading methods like skimming, scanning, or speedreading—techniques that involve sampling and attending to only a few of the available cues on the page or screen. Readers skim unfamiliar texts to get the gist of their meaning, and they skim familiar texts to confirm their familiarity with the text and to check that no new information is being presented. Scanning involves looking only for a particular letter or word, or for specific information in a text. Readers scan when they consult the phone book, look for a word in the dictionary, or hunt for a particular reference in an article, though they use scanning in these instances not because the information is already familiar but because most of it is irrelevant to their current, very focused search. Computers do this kind of reading very efficiently.

On the other hand, if readers find a text difficult because their reading skills are not fully developed, or because the text is elliptical and they lack the background knowledge to fill in the gaps, they are likely to compensate by reading redundantly. Redundant reading is like shifting into four-wheel-drive. Readers who normally read rapidly and silently are likely to slowdown and use their voices. By saying the words aloud or using inner speech to imagine the sounds of words, readers consciously employ a redundant method of comprehension—a supplement to visual recognition of words. Additional redundant reading methods include holding and reviewing the sounds of the words in echoic memory and pausing to reread phrases or sentences. A technical article from outside one’s field of expertise or a text in an unfamiliar foreign language would demand redundant reading, as would modern poetry if it is elliptical in syntax or allusion. Allusion is elliptical since it hints at other texts or ideas without being explicit. Students who have been led to believe they ought to be able to speedread any text, with full comprehension, often need to be reminded that difficult reading situa­tions call for redundant reading methods.

However, ellipsis in a text does not always require redundant reading methods if the reader has enough background knowledge. For readers who are familiar with elliptical language like abbreviations, initials, or acronyms, the ellipses become shortcuts to meaning, presenting readers with fewer words or letters to wade through—just enough to allow understanding. Classified ads, for example, often use a common vocabulary represented by abbreviations (e.g. “SWM, 37, ISO SF”), and they assume readers who are so familiar with those conventions that to use complete words would be need­lessly redundant. All abbreviations are by definition elliptical, but their ellipsis does not necessarily make reading them harder than reading the full words they stand for. Table 1 lists examples of redundancy and ellipsis in texts, in readers’ background knowledge, and in their reading methods.

Table 1

Ellipsis in Document Design

Redundancy and ellipsis can also be found in the structure of texts. Elliptical document design is characterized by a lack of orienting features, leaving readers to make their own connections between parts of the text. A text divided into many parts with no overview and few cross-references would be elliptical in design. Ellipses in hypertext design occur at the links between nodes of information if these links are not as fully or narrowly specified as in well-structured, linear texts. The links are there, but as Jay David Bolter explains, readers have more responsibility for developing coherence between the nodes they link:

Electronic mail may be another example of ellipsis in document design. As correspondents become used to the medium, they tend to eliminate some of the conventional features of letters. Salutations, opening sentences, and formal closings can be omitted for at least three reasons. First, since the level of formality in e-mail correspondence is even lower than in memos, there may be less call for polite preamble. Correspondents may feel free to plunge right into their messages. Second, information about the sender is included automatically in e-mail message headings (along with more information about the message routing than most of us need), thereby establishing some of the context. And third, due to the rapid transmission of messages there is less need to reestablish context. If one e-mail message asks a question, the reply, if prompt enough, may not need to refer to the question or enclose a copy of it, but may open directly with the answer, as if the correspondents were communicating face to face. How elliptical can an e-mail message be? One that I sent (somewhat facetiously) in response to a question was simply “‘K.”—an abbreviation of “O.K.” Then again, one might argue that some of the practices of e-mail correspondents will seem elliptical only as long as people think of these texts as letters instead of recognizing e-mail as a communication medium in its own right, with its own conventions.

As I remarked near the beginning of this discussion, conventions of syntax make it easy to identify ellipsis and redundancy at the sentence level. We can readily spot extra words or missing words. And examples of functional redundancy beyond the sentence level, like those I have men­tioned above, are easy to find because we can see the repetition on the page. I have discussed a couple of examples of ellipsis in document design, but generally speaking, ellipsis is harder to identify above the syntactic level. We can recognize a simple enthymeme (e.g. the box elder is a broadleaf tree, so it provides little shade in winter) as an elliptical form of a syllogism, and it is usually easy to infer the major premise missing from the enthymeme. But ellipsis occurs wherever readers are aware of information the writer appears to have implied but has not stated, and inferring that information at levels of abstraction much higher than the sentence level becomes harder as the possibilities multiply exponentially. What one reader reads between the lines of a text may not be the same as what another infers. And of course a text may also be elliptical in its intent. Ironic discourse, for instance, means more than it says, relying on readers to fill in the missing sense.

Ellipsis in language seems to have two main functions: first, it allows efficient reading if the reader has enough background knowledge to allow ready infererence of what has been omitted; and second, by requiring readers to make inferences, it makes the writing more engaging, more intellectually or aesthetically stimulating. Elliptical uses of language can be suggestive, denying full disclosure, inviting the reader to participate in the making of meaning. Both these functions of ellipsis can have the effect of creating a bond of respect and shared assumptions between the author and the reader. They become collaborators in the discourse. Consequently, the elliptical text is exclusive in the sense that it is not designed for all readers, just for those who can bridge the gaps within the text. And from its gaps one can infer the intended readers—those the author assumes can supply missing information or missing warrants or assumptions that support an argument. In this way, too, elliptical texts may shape their audiences, implying that they hold, or should hold, certain beliefs or knowledge.

Technical writing obviously makes use of the first function of elliptical language—to allow efficient reading—but can it take advantage as well of the second function? If we accept that technical language can be highly rhetorical, working on its readers in ways that they might not notice, then the answer must be “yes.” Evidence to support that answer can be found in Fahnestock’s analyses of scientific writing (“Connection”; “Tactics”). Writers of any kind of scientific, technical, or professional prose must routinely decide what to state and what to suppress or leave as implied assumptions. To the extent that they can, readers instinctively and obligingly fill in ellipses left by writers, whether at the syntactic or the semantic level, and the skillful writer can take advantage of that propensity. Learning to recognize ellipses in arguments, then, is an essential critical reading skill, helping readers to challenge the implications they are being led to accept.

Redundancy in Document Design

Table 1 lists examples of redundancy in text that operate at the sentence level, but it also includes some that are recognizable as good principles of docu­ment design. Most technical writing textbooks consider redundancy anath­ema, but few recognize that it operates in the engineering sense (as a form of reliability assurance) in the kind of functional repetition that characterizes the structure of many kinds of technical writing. Hendrickson’s Writing for Accountants (16-18) and Pfeiffer’s Technical Writing: A Practical Approach are two recent exceptions that show an enlightened understanding of redundancy and repetition.

While needless redundancy at the sentence level impedes busy readers, the busyness of those readers means that, faced with a lengthy report and limited time, they will probably not read the whole document cover to cover, in sequence, carefully, and at one sitting. Instead, they will likely read selectively, either following a practiced search strategy for just the informa­tion they need or engaging in more casual browsing, turning the pages in either direction, looking at the pictures, and sampling fragments of text out of sequence. The reading may also be interrupted by telephone calls, unexpected visits, appointments, and other distractions. Because this ellip­tical style of reading resembles what readers tend to do with hypertexts (Carliner 39-40), I call it “hyperactive” reading.

Documents destined to be read hyperactively under these conditions need to be designed robustly, which means that writers need to repeat and emphasize important information in several different places to ensure that the various readers with their different reading methods and purposes will all find what they need (Pfeiffer 63-64). Furthermore, multiple readers will probably each read the document in a slightly different way. The greater the number of readers and the longer the document, the more redundancy should be built into it. The trick is to try to meet the needs of all the readers while inconveniencing any one reader as little as possible, but I would suggest that in most kinds of writing it is better to include a little too much redundancy, and make some readers compensate by skimming, than to err on the side of ellipsis and make other readers work harder at interpretation. Too much redundancy is a nuisance, and if poorly designed it may obscure important points, but excessive ellipsis can be a more serious impediment for readers.

A common form of redundancy in document design is metadiscourse—passages that remind readers what they have read and announce what they are about to read. Placed prominently at the beginnings or ends of paragraphs where readers who skim are more likely to notice it, metadiscourse helps accommodate elliptical reading. Another way to ensure that elliptical readers catch the most important information in a report is to include the information in well-written section headings. These headings may be both elliptical and redundant: elliptical in syntax if they take the form of terse phrases rather than complete clauses, and redundant in the sense that the information in a heading is repeated elsewhere in the text—in the section it heads and perhaps also in a table of contents compiled from the headings in the body of the report.

A third application of functional redundancy in document design is to integrate visual aids with text. A well-integrated illustration will have an informative heading and perhaps a caption too, providing essential explana­tions for readers who look at the figures before reading the text. And if readers arrive at the illustration by way of the text, they will likely be able to interpret the illustration more efficiently if the surrounding text discusses it.

Including metadiscourse, headings, and captions is standard advice in technical writing textbooks. Holland, Charrow, and Wright, for example, recommend most of these techniques as ways of writing to multiple audienc­es, for whom they also suggest including more and less technical versions of the same information (41). My purpose in pointing to these practices as instances of redundancy in document design is to suggest a more comprehen­sive theoretical framework for them, to explain why they work in terms of what readers do. The principles of redundancy and ellipsis underlie and interrelate a variety of good writing practices.

Repeating information not only in different places in a document but also in different ways may be helpful, although opinions differ on the value of “elegant variation” (the use of synonyms to avoid repeating words). The conventional wisdom in technical writing textbooks is that elegant variation can cause confusion if readers do not realize that the different words all refer to the same thing (Brusaw, Alred, and Oliu 217; Houp and Pearsall 150). Laib concedes that elegant variation may be abused but argues that, if used well, it “becomes explication, translation, and analogy” (448).

The second point I would add is that restatement may actually contribute to the perceived conciseness of the writer. Writers tend to be most concise at places where they restate in a nutshell a point they have just explained at greater length. Alone, the “nutshell” restatement might not be developed enough to convey the point clearly, but with the support of the lengthier explanation it follows, it can provide readers with a readily portable idea. The presence of these brief, redundant restatements of main points may contrib­ute to a reader’s sense of conciseness as much as actual brevity does. Kim Sydow Campbell, for example, argues that cohesion in a text is created by repeating semantic and structural elements: “This repetition of form pro­vides a uniform background against which the semantic differences in a text are foregrounded and therefore more easily perceived” (229). An illustra­tion of Campbell’s argument might be to take a sentence like “Redundancy is characterized by repetition, while omission is the main feature of ellipsis” and revise it to include more structural repetition: “Redundancy is character­ized by repetition, while ellipsis is characterized by omission.” The similar­ities in structure and wording of the two sentences emphasize the contrast between the important words, which are different. However, it seems to me we might heighten the contrast further by keeping the structural repetition but creating ellipsis by using a single verb to serve two clauses in the same sense: “Redundancy is characterized by repetition, ellipsis by omission.

Ellipsis and Redundancy in Figures of Speech

Many figures of speech, some more flamboyant than others, operate by creating redundancy or ellipsis within or between sentences. Asyndeton is an elliptical figure of speech that uses fewer conjunctions than normal to link the items in a series: “I came, I saw, I conquered”; or, to update Caesar’s words for the modern workplace, “Megacorp follows a simple philosophy: it enters a market, it scouts the competition, it buys them out.” (While these examples of asyndeton employ ellipsis, they also use some of the structural repetition that Campbell notes. A more elliptical but less effective version of Caesar’s remarks, eliminating the repetition, would be “I came, saw, conquered.”) The figure of absolute ellipsis omits a word or words not readily inferrable from the syntax alone: “Three strikes and you’re out” (which requires a grasp of baseball to be understood).

The elliptical figure in the sentence, “Redundancy is characterized by repetition, ellipsis by omission” can be classified in George Puttenham’s scheme as zeugma—or more precisely, prozeugma (175-76). Zeugma and syllepsis are similar elliptical figures, syllepsis being a construction in which one word seems to apply equally to two others but does not. Syllepsis may involve a grammatical anomaly, as in this example of a singular verb serving both singular and plural subjects: “The larger screw threads counterclock­wise, all the others clockwise”; or it may be semantic, as in this example of metaphorical and literal expressions branching from the same verb: “In Las Vegas he lost his head and his inheritance.” Puttenham makes a clear contrast between syllepsis, whether grammatical or metaphorical, and zeugma, which he defines as an elliptical figure involving neither grammatical nor metaphorical play. For Puttenham, a zeugma could simply be a construction where two or more subjects share a single predicate: “Fellowes and friends and kinne forsooke me quite” (175). However, since Puttenham’s day, as the Oxford English Dictionary notes, syllepsis and zeugma have become confused and zeugma commonly denotes a metaphorical construction. For instance, Webster’s II New Riverside University Dictionary (Houghton Mifflin, 1984) defines both terms as having metaphorical elements, and offers very similar examples to illustrate syllepsis (“We lost our coats and our tempers”) and zeugma (“We changed our minds and our clothes”). On the other hand, the 1993 edition of Merriam Webster’s College Dictionary recognizes the distinc­tion Puttenham makes between grammatical and metaphorical syllepsis but equates zeugma with the metaphorical kind of syllepsis (“She opened her door and her heart to the lonely boy”).

This shift in the definition of zeugma might seem but a footnote to the evolution of the English language, a splitting of etymological hairs that only a cloistered rhetorician could love; however, I believe that this shift actually reflects a telling and regrettable shift in attitudes towards figures of speech. We tend now to make a clear but perhaps simplistic distinction between figurative language and plain language. Figures of speech are thought to belong only in “creative” writing and to have no place in professional or technical writing. One problem with this belief is that metaphorical lan­guage, used carefully, can be a powerful tool in technical writing, as some scholars have shown (see Colby; Hoffman; Plung). A striking example of metaphor appeared in a paper about the effects of various kinds of damage on the strength of ships’ hulls. The paper includes a photograph showing how water pressure can cause hull plates to become concave between the ribs of a ship’s frame. The author calls these indentations “‘hungry horse’ deforma­tions” (C. S. Smith 97). This metaphor is both elliptical and redundant: elliptical in the sense that in just two words it evokes a comparison between a ship’s hull and the image of an undernourished animal whose ribs are showing through its skin—an image the author assumed all readers could supply—and redundant in the sense that it supplements the technical expla­nation of the deformations. The net effect is to give readers a memorable image of the phenomenon.

One could argue the “hungry horse” example is an exception proving the rule that figurative language does not usually belong in workplace discourse, and I would certainly not recommend we teach our technical writing students to ornament their papers with masses of purple metaphors. However—and this is a second problem with the belief that figurative language belongs only in creative writing—not all figurative language is metaphorical. A good deal of figurative language is based on redundancy or ellipsis, characteristics that can be found in any genre. In other words, rather than seeing figurative language as a particular and quite different kind of language that should be taught only in literature or creative writing classes, I believe we should show our technical writing students—and indeed all our writing students—how redundancy and ellipsis can be manipulated to produce figurative language, some of it metaphorical, some of it not, some of it appropriate for technical writing, some of it not. If we broaden our view of figurative language in this way, we may find that workplace writing includes not just the occasional metaphor but also its more reticent cousins, the utilitarian figurative ele­ments like listing (which I will discuss further below) that are important, everyday tools of professional writers. Rather than accepting the traditional divide between creative and technical writing, I prefer to see a continuum between them, recognizing that they share some common skills and that the conventions of technical writing include figurative elements.

Some figures of speech create redundancy by repetition or addition. Polysyndeton uses more conjunctions than necessary to link the items in a series, as in the familiar inscription on the main Manhattan Post Office: “Neither snow, nor rain, nor heat, nor gloom of night stays these couriers from the swift completion of their appointed rounds.” Epistrophe repeats the same word or phrase at the end of each clause in a sequence, as in this passage from the Bible: “When I was a child I spake as a child, I understood as a child, I thought as a child: but when I became a man, I put away childish things” (1 Cor. 13.11). Anaphora repeats the same word or phrase at the start of each sentence in a passage, as in this excerpt from Winston Churchill’s “Dunkirk” speech demonstrates:

My second point about the revision is that while changing redundancy to ellipsis purged the emotion from the original, it did not eliminate the figurative language of the original. It simply changed it from a redundant figure (anaphora) to an elliptical one (something close to zeugma, as Puttenham defines it). I would suggest we might consider this kind of list as an example of a utilitarian figure of speech.

Conclusion

When we teach students to recognize redundancy and ellipsis in what they read and write, we should not teach them to shun either instinctively. Rather, we should try to teach them the difference between functional redundancy and needless redundancy, between expandable prose and expendable prose, between ellipsis that engages readers and ellipsis that alienates them. Ex­treme redundancy or ellipsis is not necessarily excessive; it depends entirely on the rhetorical situation—who is reading, what is being written, and what needs to be emphasized. Laib argues that we should teach “not a prescrip­tively balanced style but proportion, equity, and dialectic in the values we teach, so that in their own writing students can strike a balance that fits the occasion” (457). I would argue we should not teach students to ask them­selves, “Is there redundancy in my writing and how can I get rid of it?” but rather, “Have I expanded or condensed my writing and repeated information at the appropriate places for my readers?”

Finding the right proportions of redundancy and ellipsis for a given topic and a more or less given (but not entirely known) combination of readers is a challenge. I have found myself attending to it at all stages of composing this paper, right through the final editorial changes. Comments from anonymous reviewers on earlier drafts helped me reinforce certain load-bearing passages and pare down others where I learned I could count on my readers to bear some of the burden of comprehension. Adding and removing words is a fundamental activity of revision, but writing instruction seldom provides students with a coherent theory of when and why to add or remove. Explain­ing redundancy and ellipsis in terms of what readers do in response to those features contributes to such a theory. It does not make writing any simpler or easier, but it does offer writers a powerful principle for revising.