Improving Academic Writing

Jonathan Bennett and Samuel Gorovitz

From: Teaching Philosophy 20:2, June 1997

Introduction

Graduate students typically aspire to write the sort of prose they read in the leading journals in their disciplines; faculty encourage this aspiration and largely determine which journals the students take as models. But too much academic writing, even in prestigious journals and books, is ‘academic’ in the worst sense. It labors ponderously, inviting the reader to think of wading upstream against treacle. The point is general, but as philosophers we will emphasize problems and propose remedies in our own discipline.

As Patricia Limerick has noted, ‘The politically correct and the politically incorrect come together in the violence they commit against the English language. . . . Everyone knows that today’s college students cannot write, but few seem willing to admit that the faculty who denounce them are not doing much better. . . . It is, in truth, difficult to persuade students to write well when they find so few good examples in their assigned reading.’¹

This problem is underscored by the recent citation of award winners in a bad academic prose contest, selected by the editors of an academic journal. Their choices of ‘the ugliest, most stylistically awful’ prose, drawn from books published by reputable houses and from articles in reputable journals, are grotesque as models of writing that graduate students might be tempted to emulate.² This sorry state of affairs is tolerated and perhaps even encouraged by typical graduate program faculty. As Limerick observes, ‘Graduate school implants in many people the belief that there are terrible penalties to be paid for writing clearly.’

Philosophy, like all fields, includes some fine writing. Even when it is admired for its expository quality, however, it tends to be studied only for its substantive content. The works of Russell and Quine, for example, are commonly read for their philosophical importance, but rarely analyzed with students for the crafting of the prose.

Yet writing well matters—morally, prudentially, and philosophically. It matters morally because in writing well one is being kind to one’s readers. It matters prudentially because in writing well one increases the chance of having readers—and career success and personal gratification may depend on

that. And it matters philosophically. Often when wrestling with a stylistic problem—how to avoid a bothersome repetition, clarify a change in terminology, or simplify a long and boring paragraph, for example—one finds that the source of trouble was some specific failure in philosophical thinking. This seems just magical, but it is real.

Despite our sympathy for Jack Point’s views, we do not advocate sugar-coated lilies in professional writing. We do believe it important to help our graduate students and junior colleagues to develop a commitment to clear, effective writing and a confident sense that better writing is possible for anyone who is serious about it. Here, we offer to them and those who train them some advice, born of our own experience, on how to improve academic writing.

Improving Sentences

Many small improvements in a text are possible with relatively little effort, well worth investing for the sake of more polished prose. Even serious writing problems are not just important and difficult; they are largely soluble with careful work.

Jonathan Bennett observes:

My first drafts tend to be dreadful, but my published work is much better. In between, I do surgery on my prose. I am increasingly impressed by how much improvement I can make just by applying straightforward rules or procedures that can be stated, taught and learned. Years of practice help one make better use of the rules, but their core—and much of the good they can do—is immediately accessible. Revisions that improve writing typically also shorten it. Although clarity is our principal concern here, we also discuss the virtues of brevity below. In most of our illustrations, the suggested revisions both clarify and shorten; word lengths are indicated in brackets in some of the examples to make the effect explicit.

The ‘Bennett Rules’ are these:

Verbs are better than nouns 

Compare: 

There is a difference between x and y. 

x differs from y.

Adverbs are better than adjectives 

He is a clear writer.

He writes clearly.

In both examples the structurally necessary but overused verb ‘to be’ drops out; it often does when prose is improved. Minimizing its use has the additional advantage of eliminating passive constructions, each instance of which should be challenged, and which are almost always lamentable because of their imprecision. Compare:

The patient was observed to be agitated.

Dr. Notewell observed that the patient was agitated. Verbs can also eliminate the kind of nominalization that bloats writing. Compare:

We are in agreement that. . . We agree that. . .

And

He adduced an argument for the proposition that . . .

He argued that. . .

A longer example from a journal article: Once it is acknowledged that genuine verdicts of conscience may be in conflict, the theory must be abandoned that the power from which those verdicts proceed is distinct from reason and superior to it. And the only plausible alternative is that put forward by Aquinas: that the verdicts of conscience derive from a disposition of ordinary human reason. They can be in conflict with one another because they can err; and there are two possible sources of such error. First, the disposition from which they derive, which

Aquinas called synderesis, is not developed equally in all human beings. Everybody has some understanding of the principles of morality and, inasmuch as he has it, cannot be mistaken about them. In this sense, synderesis cannot err. But it can fail; that is, a man’s understanding of the principles may be limited, and naturally, when he judges what lies beyond the limits of his understanding, the verdicts of his conscience will be defective. Second, even when he does understand the principles applying to a given case, that application often calls for an act of subsumption, in which the case is brought under the principle by one or more specificatory premises. Although this process is not as a rule explicit, especially when the conscience in question is uninstructed, the verdicts of conscience nevertheless presuppose it. And error is possible about specificatory premises, as it is about any outcome of nontrivial conceptual analysis. (237 words)

The revision of that began with a computer search for all forms of the verb ‘to be’:

Because genuine verdicts of conscience may conflict, the power from which those verdicts proceed cannot be distinct from reason and superior to it. The only plausible alternative view is that of Aquinas: the verdicts of conscience derive from a disposition of ordinary human reason. They can conflict with one another because they can err, for either of two reasons. First, the disposition from which they derive, which

Aquinas called synderesis, is not developed equally in all human beings. Everybody has some understanding of the principles of morality and, inasmuch as he has it, cannot be mistaken about them. In this sense, synderesis cannot err. But it can fail, because a man’s understanding of the principles may be limited, so that when he judges what lies beyond the limits of his understanding the verdicts of his conscience will be defective. Second, even when he understands the principles that apply to a given case, applying them may require an act of subsumption in which he brings the case—though perhaps not explicitly—under the principle by one or more specificatory premises. There can be error about specificatory premises, as about any outcome of nontrivial conceptual analysis. (195 words)

Favor the Anglo-Saxon. English has Teutonic as well as Latin roots; for reasons having to do with the social structures in England after the Norman conquest, the Latin-French side of the language is prominent in learned and abstract discourse, while the simpler Anglo-Saxon side has survived for more down to earth matters. It is a great merit in English prose to have a good mix of the two, but philosophers have to work for this because they are subject to a magnetic pull from the Latin-French side.

Some potent examples:

(a)   ‘possible’, ‘impossible’, ‘necessary’, etc. can often give place to ‘can’, ‘cannot’, ‘must’, etc.

Compare:

Is it possible for him to. . . ? Can he . . .?

(b)   replace ‘at a subsequent time’ by ‘later’, ‘at every time’ by ‘always’, and so on. Compare:

She wrote it a century prior to the present time. She wrote it a century ago.

(c)    ‘condition’ especially when combined with ‘necessary’ or ‘sufficient’, does work that can be nicely done with ‘if’, ‘only if’, ‘unless’, ‘not unless’, and so on.

(d)   Compare:

He has an obligation to pay the bill. He ought to pay the bill.

Banish ‘very’ and its ilk. Find each occurrence of ‘very’, ‘extremely’, ‘really’, ‘quite’ and so on; remove almost all of them. This greatly strengthens the work.

Abstract nouns should be fought like the devil, especially by philosophers. This is among the reasons for favoring the Anglo-Saxon. Because we so often need abstract terms

(which typically come from Latin or Greek), we should avoid them where we can. This is from a reputable journal: A more abstract exploration of the formulations’ implications for each other, however, will reveal that each is unacceptable.

Of five consecutive words, three end in ‘ation’; it looks terrible, and is dreary to read. Compare: If we explore more abstractly what each formulation implies for the others, we see that none is acceptable. Not great, but it is better. Before submitting anything for publication, Bennett also asks the software to reveal every instance of ‘ation’, ‘ness’, ‘ism’, ‘ility’ and their plurals—and asks of each, ‘Is this earning its keep?’ Often the answer is

‘No.’

Another published example:

None of these positions would make the distributive hybrid vulnerable to the objection to utilitarianism dealing with distributive justice. For that objection, as I have said, arises ultimately in response to the utilitarian conception of the overall good, which ranks states of affairs according to the amount of total satisfaction they contain. And the distributive hybrid’s rejection of that conception of the overall good is unequivocal. Its institutional principles, whatever they may be, rely on the distributive principle for ranking overall states of affairs. Thus none of the possible institutional principles would require that some people’s life prospects be sacrificed in order to increase the non-essential satisfactions of other people whenever that would serve to maximize total aggregate satisfaction. For they all reject the conception of the overall good which leads utilitarianism to require just that. [137 words] A revision:

None of these positions would expose the distributive hybrid to the objection to utilitarianism dealing with distributive justice. For that, as I have said, objects basically to the conception of the overall good which ranks states of affairs according to how much satisfaction they contain; and the distributive hybrid flatly rejects that conception. Its institutional principles, whatever they may be, are guided by the distributive principle in their ranking of overall states of affairs, so they cannot require that some people’s life prospects be sacrificed so as to increase the non-essential satisfactions of other people whenever that would maximize total aggregate satisfaction. [103 words]

Avoid undue repetition This is a prominent type of bad writing in contemporary philosophy. There are two main ways of avoiding it: (i) With pronouns, (ii) by letting an expression operate more than once without being explicitly repeated. Here is an example from the literature:

Classical utilitarianism, which ranks states of affairs according to the amount of total satisfaction they contain, is the most familiar consequentialist view. But classical utilitarianism is widely thought too crude a theory. Although its defenders point with approval to its simplicity, critics charge that this simplicity is achieved at too high a cost. They argue that utilitarianism relies on implausible assumptions about human motivation, incorporates a strained and superficial view of the human good, and ignores a host of important considerations about justice, fairness, and the character of human agency. [92 words] Suggested repair:

Classical utilitarianism, which ranks states of affairs by how much satisfaction they contain, is the most familiar consequentialist view. It is widely judged to be too crude a theory. Its simplicity, to which its defenders point with approval, is said by critics to come at too high a cost. Utilitarianism, they argue, relies on implausible assumptions about human motivation, takes a strained and superficial view of human good, and ignores much of what matters about justice, fairness, and purposeful human action. [81 words]

This repair has the following effect on word occurrences:

‘utilitarianism’ from 3 to 2; ‘simplicity’ from 2 to 1; ‘considerations’ eliminated; ‘incorporate’ eliminated.

Be careful with commas. Use a comma where it is needed, or even optional but clearly helpful. Avoid it otherwise. G. E. Moore wrote ‘There is, in this, a confusion, with which, however, we need not deal’. (Principia Ethica, p. 65.) Compare: ‘There is a confusion in this with which, however, we need not deal.’ A recent translation of Leibniz offers:

I respond that the reason is the same in both cases and for both sorts of things, namely, for all changes, of both spiritual and material things, there is a place, so to speak, in the order of succession, that is, in time, and for all changes, of both spiritual and material things, there is a place in the order of coexistents, that is, in space. [66 words]

The following version reduces the thirteen commas to five:

I respond that the reason is the same in both cases and for both sorts of things: namely that for all changes of both spiritual and material things there is a place, so to speak, in the order of succession, i.e. in time, and a place in the order of coexistents, i.e. in space. [52 words]

Attend to the sound. Before regarding a piece of writing as finished, test it by ear. Bennett wrote to a graduate student:

Gilbert Ryle once told me, ‘What doesn’t read well to the ear doesn’t read well to the eye’, and that changed my life. More than any other one thing, that insight showed me how to start climbing out of the garbage pit up onto the plain of decent prose. I had some of my material read to me while I listened with my eyes closed, was appalled by how ugly and boring it was, and took action. I no longer use exactly this technique, but I read aloud to myself everything I write for publication. I recently sat in my study at home and read my book The Act Itself aloud in ringing tones, imagining an audience and aiming to do the performance with gusto. Whenever my confidence ebbed and my voice wavered because the prose was not moving properly, I rewrote. For your first few professional years, though, I urge you to submit yourself to the discipline of listening to your own prose without at the same time following it on the page. It may well be the worst experience of your intellectual life. If so, it will also be one of the most valuable.

Here are two examples employing several of the above principles:

Original:

If what has been previously claimed (though not conclusively defended) is true—that the thesis that ‘ought’ implies ‘can’ is untenable even in the moral domain— then it is altogether possible that a man believe that he ought to act in such and such a way and yet, in the relevant circumstances, fail to act not merely through reasons of ignorance, inadvertence, change of heart, or the like but also (a decisive possibility) because of an incapacity to act. [79 words]

Rewritten:

If I have been right in saying (not proving) that “‘Ought” implies “can”’ is untenable even in the moral domain, then someone can fail to act as he thinks he ought because he cannot act in that way. [37 words]

In simplifying the prose, one achieves both greater clarity and brevity; the revision is less than half the length of the original. Yet no valuable substance is lost. Note also the removal of the confusing nested ‘that P’ formulation—‘that the thesis that “ought” implies “can” is’ etc. In general, it is good to minimize the use of that form.

Another original passage:

What emotivism is against is not that ethical judgments might be made to follow from factual judgments (for this is precisely what persuasive definitions achieve, and there is nothing essentially wrong with such definitions); rather, it is against the reductionist thesis that ethics can be intellectually respectable only if ethical judgments can be analyzed as following the expression-of-belief model appropriate for scientific statements, so that such judgments are ‘really’ cognitive, and only if ethical agreements can be analyzed as following the disagreement-in-belief model appropriate for science. [90 words]

Revised:

Emotivism does not oppose making ethical judgments follow from factual ones; that is what persuasive definitions do, and nothing need be wrong with them. Rather, it opposes the reductionist thesis that for ethics to be intellectually respectable its judgments must follow the expression-of-belief model appropriate for scientific statements, so that they are ‘really’ cognitive, and disagreements about them must follow the disagreement-in-belief model appropriate for science. [70 words]

More Global Improvements

Use technical notation cautiously Unnecessary symbolic machinery detracts from clarity. Formal logic can help in the development of an argument without being displayed on the page. Be clear about where the quantifiers would come and what they would be, and where the brackets and operators would occur, if you did express your point in formal notation.

But do not use it in print without good reason. It does not add rigor, clarity, or even the appearance of either, unless it is present because it is needed.

Use any formal apparatus judiciously. If you refer to items by letters of the alphabet, know why. When possible, use a name with mnemonic value. E.g. in contrasting two imaginary people who disagree about whether it would be right for an agent to perform a specific action in a given situation, you may need names for them. You could use ‘X’ and ‘Y’ or genuine names such as ‘Jack’ and ‘Zack.’ Why not use names that help the reader remember which is which, e.g. by calling the person who thinks the action would be right ‘Rebecca’ and the one who thinks it would be wrong ‘Wendy’?

Question acronyms. When using an acronym, ask why. Why not have the phrase it abbreviates written in full each time? If that would add unacceptably to the length, are you using it too much? Having asked these questions, you may rightly conclude that the acronym is justified; if so, explain it fully when it first appears.

Attend to problems of order. If you refer forward to something before you are ready to explain it fully, it is likely that a mistake led you to do so. This is also likely when you assert something and later retract it. Figure out a better way of doing things: the better way will always involve a substantive improvement. Problems in the order of exposition are never purely stylistic.

Avoid prose in footnotes. When you are tempted to put prose in a footnote, ask why you want the reader to jump down the page and up again. What good can this do? If you answer ‘The note is to be read when the reader reaches the end of the page, or it need not be read at all’, how is the reader to know that? Do you honestly regard the footnotes as dispensable? The concern that one tries to resolve through substantive footnotes can often be eased by the judicious use of parentheses in the main text. One can elegantly end a paragraph with a sentence or two in parentheses; the reader knows when to deal with that, and is advised that it is not part of the main thread. In many cases, however, material in a footnote can be smoothly woven into the text without the use of parentheses. Not infrequently, the obstacle to doing that (which led initially to placing the material in a footnote) is that there is some philosophical matter that has not been thought out thoroughly enough. Sometimes, of course, material in a footnote is better dropped altogether.

We realize that this advice is contrary to established practice in some disciplines—most especially law, in which the space devoted to footnotes sometimes exceeds that in the body of the text. Even there, however, a well-written article can be read through without interruption by the footnotes, to which the reader can return later if more detail is wanted.

Read with charity; report with more charity. Make your opponent’s case look as good as possible; try to leave it in a more cogent and convincing form than you found it in. This valuable courtesy reduces time wasted in allegations of injustice and misrepresentation and it conduces to philosophical discovery. Perhaps your opponent is right after all or not wrong in the way you had thought. If on your interpretation a writer has said something plainly false or obviously stupid, then assume you have misinterpreted it. If in a given case you are positive that this rule doesn’t apply, ask why you are spending time discussing obvious falsehood or idiocy. Avoid sneering language.

Prepare abstracts. Before submitting for publication a long article or a book, write a version of it that is five or ten percent as long, omitting much detail but preserving the main thread and over-all shape. If your work cannot be abridged in that way, there is something wrong with it—some unacknowledged problem of form, order, or over-all coherence. Find and fix that trouble, and then try again to write the abridged version. This arduous and time-consuming work is justified by the great rewards that it brings.

Deliberate about divisions. Breaks in the text, dividing lines, numbered sections, headings, and chapter titles are part of what helps the reader understand the author’s meaning. They merit care and deliberation.

Index your own work. For any book that you write, make the index yourself; there is no chance that anyone else will do it satisfactorily. A computer-generated index is at best a crude first step: many of its entries will have to be deleted, and—because a good index is based on the occurrence of concepts, not of words—many others will need to be added. Do not let a single entry be too long. If you have more than about ten page-numbers for a given entry, consider dividing it into sub-entries.

Putting It Another Way

The capacity to put the same point differently is crucial to the ability to improve one’s writing through revision. But how can one become more adept at putting things differently? Like so many other skills, this one can be enriched by exercise and practice. It helps greatly to have an appreciative, creative, and even whimsical sense of the language. That can be fostered by the kinds of games that language lovers play. Doing crossword puzzles, writing palindromes, finding anagrams, and the like, thus have practical benefit. It can even be fostered by the kinds of language games one plays with children, such as Rhymes of Opposites (I say ‘hot’, you reply ‘bold’) and Opposites of Rhymes (I say ‘bold’, you say ‘young’).

Writing is greatly enriched by reading good literature and thinking about why it is well-written. One’s ear for language will be more sensitive if one regularly reads good prose stylists: Melville, Twain, Dickens, Jane Austen, George Eliot, Tolstoy, Henry James. And they ought not just to be classical authors. More recent or contemporary writers should be in the mix: E. B. White, James Thurber, John McPhee, Tobias Wolff, Calvin Trillin, Maeve Binchy, Michael Frayn, Alice Munro, Amy Tan, Robert Butler, Julia Alvarez, Jonathan Raban, Richard Rodriquez—pick your own most admired writers. Reading such fiction and essays can deepen and enrich philosophical work in other ways as well.

More whimsical explorations of the power and capacity of language are also valuable. Rarely do we write under extreme constraints. Rather, we express what we have to say as it comes to us; then, we work to improve what we have said by critically reassessing it. But even if we are reasonably adept at finding other, better ways to say it, we are likely to be limited by the language we first used, and unlikely to move far from it, if we are not creative at seeing how else we might have said it. Only with an imaginative sense of the language are we able to find entirely new ways to present what we may at first have said with too little clarity or grace. One example is a dazzling exposition of the Second Incompleteness Theorem, presented by the late George Boolos in 1994 to a diverse audience entirely in words of one syllable. Wanting to emphasize the accessibility of the basic ideas to a general audience, Boolos sought to distill the key issues—and an account of why they matter—until they were expressed simply enough to be understood by intelligent listeners with no relevant background. Summarizing the main point, he said:

In fact, if math is not a lot of bunk, then no claim of the form ‘claim X can’t be proved’ can be proved.¹

Too many extended philosophical discussions neglect to speak to the question of why the argument matters; this neglect limits the appeal of the writing to readers who already understand the point. Explaining the significance simply, however, can expand the audience. It need not be done in words of one syllable, but Boolos’s example is worth bearing in mind as a rare model of what can be done.

It is always gratifying to see a passage improve as the result of one’s labors, but sometimes more than satisfaction is at issue. Samuel Gorovitz notes:

Having written dozens of articles in philosophy journals and books, I had a transforming experience when given an opportunity to write newspaper editorials. Strictly limited to 800 words, I had to learn to make every character count. This was entirely new. Always, in such writing, my first drafts are too long—1,000 or 1,200 words. I do not want to relinquish content—and certainly do not want an editor to make cuts in my behalf—so I must find shorter ways of saying the same things. It sometimes takes four or five drafts. The result is always better writing, and little content is ever lost.

Efficient writing has greater clarity and precision, and also pleases journal and book editors because it does not waste valuable space and can be published more economically. Editors do not admire flabby writing, but linguistic flab abounds. One way to reduce it is to banish phrases that add nothing to a text, such as ‘It is important to note that. . . ’, ‘The fact of the matter is that. . . ’, ‘I believe that. . . ’, ‘I think that. . . ’, ‘I feel that. . . ’, ‘In fact. . . ’. Such linguistic fillers may almost always be deleted to advantage. Once one develops the goal and habit of deleting the superfluous, detecting verbosity becomes much easier. An efficient writer will not say—nor will a careful editor allow—‘I would like to thank my mentor, Minnie Muse. . . ’, rather than ‘I thank my mentor, Minnie Muse. . . ’.

Many offenses other than verbosity displease editors and referees; they all should be avoided. As frequent referees for professional journals and publishers, we are utterly dismayed at how often we are asked to assess manuscripts that are badly written—often in surprisingly crude ways. Problems of grammar, spelling, punctuation, word choice, consistency of usage, and the like are all too common, and they weigh—sometimes heavily or even decisively—against a favorable reading of any manuscript.

Careful proofreading is essential to the process of writing well, yet is an art too little honored by many writers. With the advent of spell checkers, we now see prose each word of which is correctly spelled, but many words of which are obviously not the intended words. Only a meticulous and labored attention to detail will filter out all the errors. Manuscripts are often submitted for publication without that attention, yet an error in even a single character can make a substantive difference. Even if a submitted text is perfect, it must be reviewed later with the same scrutiny to detect errors introduced after acceptance. For example, for an editorial on criminal justice, Samuel Gorovitz wrote:

The citizens of other countries often see us as living amid violence and chaos. They know that we shoot one another an order of magnitude more often than Europeans do. . . .

But what appeared in print was:

The citizens of other countries often see us as living amid violence and chaos. They know that we shoot one another in order of magnitude more often than Europeans do. . . ¹

This suggests that our homicides proceed according to height or weight.

More embarrassing—or amusing—is this passage, about the New York Catholic Conference’s opposition—based on a sense of the potential value of human suffering—to legislation granting status to health care proxies. The intended text was:

The political resistance in New York had the support of at least one conservative religious leader. Cardinal John J. O’Connor, writing in Catholic New York (July 20, 1989), explained why he refrained from supporting the bill. . . .

But what appeared was:

The political resistance in New York had the support of at least one conservative religious leader. Cardinal John J. O’Connor, writhing in Catholic New York (July 20, 1989), explained why he refrained from supporting the bill. . . .²

This error survived the author’s proofreading through all stages from first draft to page proofs, but when the book was in print it leapt from the page. The copyeditor, of course, had no way to know the wording was not intended. Neither copyeditor nor spell-checker should be expected to distinguish, say, ‘the quality of life’ from ‘the duality of life’, or ‘the justification of belief’ from ‘the justification of relief’. There is no substitute for exquisite care on the part of the author.

Teaching Tomorrow’s Teachers

The quality of our students’ writing is not just influenced by what they read in the professional literature. Our sense is that they are especially influenced by the way their own teachers write and by how we read, analyze, and discuss what we assign. If we only look to the substance, we encourage our students to undervalue the importance of the quality of writing. But if we make that important to our inquiry, we encourage a concern with expository quality. This is so not only regarding what we read together with our students, but in how we react to the writing they submit to us. If we stress that good writing is important and show how it can be achieved by deliberate and structured effort, we can reduce our own suffering when we read our students’ work and can improve their expository skills.

For those of us who teach in doctoral programs, this purpose can also be advanced as we mentor graduate students in their capacity as teaching assistants. For example, in a large undergraduate course, Gorovitz sought to maintain a serious focus on the students’ writing despite the class size of nearly 400. Six teaching assistants met with students in small weekly discussion sections, of which there were sixteen.

Student essays were to be graded by the TAs. But TAs might not comment on the papers in a uniform, sufficiently detailed, and appropriate way, and a student’s grade might then depend on which TA he or she had. These concerns provided a context for focussing graduate student attention on their own writing as well as on that of their undergraduate students.

At the start of the course, each TA submitted a writing sample, which was quickly returned with comments. When the first undergraduate essay was submitted, photocopies of six randomly selected papers were distributed to the TAs. An additional paper was graded by the professor, with copies of the result distributed to the six TAs as a model. The TAs then each graded the other six papers. Those grades, arrayed at a course staff meeting early in the term, showed great diversity with individual papers receiving grades ranging from A- to Cin one case and from C+ to F in another.

Each TA then had to defend his or her grading choices, specific comments on the essays, and the absence of comments thought by others to be important. The discussions covered the rank ordering of student papers, the standards for grading, matters of substance and style, and the like. Each paper was discussed until agreement was reached about what its grade should be and what comments should be on the paper. Then, the TAs graded additional papers, submitting their grading work for further review before being authorized to grade papers on their own.

Quality of writing was only one factor in this process, but one emphasized as important for quality of thought and substance. (This emphasis initially surprised some of the TAs.) The TAs, held to account for meticulous attention to the details of writing by their students, became more aware of their own styles as writers and more explicitly reflective about what contributes to or detracts from admirable prose. With or without assistants, the teaching of writing is labor intensive. The challenge is especially daunting for those who have large classes and little or no help. No matter how clearly any set of guidelines is fashioned, there is no substitute for the tedious task of subjecting students’ work to careful line by line attention that helps them see not only that revision is needed, but of what sort, and helps them learn to make such improvements for themselves. Too often that labor by faculty is not only unrewarded, but punished. The professor who strives to improve students’ writing rarely gets extra professional recognition for doing so, and is sometimes even criticized for being less productive than had the time been spent on original scholarly work or more visible pedagogical efforts. Efficient approaches to the task are therefore especially welcome.

Some faculty have student work submitted on disk or by email, allowing the professor to intersperse commentary directly into the text. Asking students to submit audio cassettes along with their papers, and providing detailed spoken commentary, can also be effective. An advantage this has over a conference is that students can, and do, listen repeatedly to such tapes, whereas what they are told in a meeting can be largely ephemeral, even if they take notes. If modest funds are available, it can help to hire a superb student—even an unusually qualified undergraduate—to review papers for the technical accuracy of the writing before they are read for content by the faculty member. In some contexts, it can be useful to have students review each others’ work. These and other methods can help greatly with undergraduate students, and if we work with graduate students the use of these methods both helps prepare them for their own teaching careers and reinforces the importance to them of the quality of their own writing.

Concluding Remarks

Fine writing is a reward for serious labor. Most of us can write well only with self-conscious, sustained effort. And that effort rewards most when it is informed by imagination, scrupulous attention to detail, a sense of the sound and resonance of language, a love of well-crafted prose of many styles, and the ability and dedication to step back from the task of making our point (and getting it right) to devote comparable care to getting it across.

John Updike illustrates this viewpoint beautifully in his discussion of Lincoln’s sense of language:

The Lincoln touch can be seen in the last paragraph of the First Inaugural Address, which was proposed by Seward—who felt Lincoln’s original ending to be too militant—in this form:

We are not we must not be aliens or enemies but fellow countrymen and brethren. . . . The mystic chords which proceeding from so many battle fields and so many patriot graves pass through all the hearts and all the hearths in this broad continent of ours will yet again harmonize in their ancient music when breathed upon by the guardian angel of the nation.

Lincoln made this of it:

I am loth to close. We are not enemies, but friends. We must not be enemies. . . . The mystic chords of memory, stretching from every battlefield, and patriot grave, to every living heart and hearthstone, all over this broad land, will yet swell the chorus of the Union, when again touched, as surely they will be, by the better angels of our nature.

A speaking intimacy and, with it, a compelling urgency and warmth were achieved through simplification and a subtly improved concreteness. ‘Fellow country-men and brethren’ became ‘friends’,  ‘hearths’ became ‘hearthstones’, and Seward’s forced ‘mystic chords. . . patriot graves. . . hearts. . . hearths. . . ancient music. . . guardian angel’ trope was deftly broken into separate components; especially striking is the way in which the upward glance at the supposed ‘guardian angel of the nation’ was turned inward, to the ‘better angels of our nature.’ Like Twain, Lincoln pared Latinate rotundity from American English and conjured music from plain words.¹

Understanding such matters is not a function of innate ability or a natural linguistic grace that some of us have and others irremediably lack. To a large extent, we can choose how well we write and we can influence how well our students write. Of course, literary talent is not mechanical; creative writing is an art even more than a craft. But clear, accurate prose should be within reach for anyone with enough worth saying to appear in print.

Although we focus here on aspects of writing well within our own discipline, our observations and suggestions may be of use to other disciplines as well. We have no illusion of having said anything comprehensive, or that our advice, if taken to heart, will eliminate flawed writing. We realize all too well that much of our own writing could have benefitted from this advice, and consider ourselves to be still learning how to write better—and still working hard at it.

Jonathan Bennett and Samuel Gorovitz

Syracuse University