The following is a list of solutions to problems I have encountered repeatedly in my students’ formal writing, such as coursework, research papers, and literature surveys.
It is a long list. People have a lot of problems.
Some of the items sound picky or trivial, even to me. Yet bad grammar, bad style, and poor organization will make it very difficult for you to convey your ideas clearly and professionally, and will limit your academic and professional success. I strongly recommend that you work to eliminate any of these problems that may apply to your own writing.
— Dr. James A. Bednar
Rules for formal writing are quite strict, though often unstated. Formal writing is used in academic and scientific settings whenever you want to convey your ideas to a wide audience, with many possible backgrounds and assumptions. Unlike casual conversation or emails to friends, formal writing needs to be clear, unambiguous, literal, and well structured.
Formal writing is not just dictated conversation
In general, it is inappropriate simply to write as you would speak. In conversation, the listener can ask for clarification or elaboration easily, and thus the speaker can use imprecise language, ramble from topic to topic freely, and so on. Formal writing must instead stand on its own, conveying the author’s thesis clearly through words alone. As a result, formal writing requires substantial effort to construct meaningful sentences, paragraphs, and arguments relevant to a well-defined thesis. The best formal writing will be difficult to write but very easy to read. The author’s time and effort spent on writing will be repaid with the time and effort saved by the (many) readers.
Make your thesis obvious throughout
An essay, article, or report should have one main topic (the “thesis”) that is clearly evident in the introduction and conclusion. Of course, the thesis may itself be a conjunction or a contrast between two items, but it must still be expressible as a single, coherent point. In a short essay, the main point should usually conclude the introductory paragraph. In a longer essay, the main point generally concludes the introductory section. The reader should never be in any doubt about what your thesis is; whenever you think it might not be absolutely obvious, remind the reader again.
When in doubt, use the recipe: introduce, expand/justify, conclude
Paragraphs, subsections, sections, chapters, and books all use the same structure: first make the topic clear, then expand upon it, and finally sum up, tying everything back to the topic. At each level, you need to tell the reader what you will be trying to say (in this paragraph, section, etc.), then you need to cover all the relevant material, clearly relating it to your stated point, and finally you need to tie the subtopics together so that they do indeed add up to establish the point that you promised.
Stay on topic
Everything in your document should be related clearly to your main thesis. You can write other papers later for anything else you might want to say. The reason your reader is reading this particular paper of yours is that he or she wants to know about your main topic, not simply about everything you might want to say (unless for some narcissistic reason “everything you might want to say” is your clearly stated main topic).
Conversely, there is no need to bring up items simply because they relate to your main topic, if you do not have anything to say about them. If you do bring something up, say something important about it!
Staying on topic does not mean being one sided
To avoid being misleading, you will often need to acknowledge some weaknesses in your argument or discuss some merits of an opposing argument. It is quite appropriate to discuss such opposing views when they are relevant, i.e., when they relate directly to the main topic of your paper. For instance, if you are reviewing a paper and arguing that it was not written well overall, it is usually a good idea to point out the few things that were done well, e.g. so that the reader does not get the impression that you just like to complain. Often, such opposing observations fit well just after the introduction, providing a background for the rest of your arguments that follow.
Whenever you do include such material, i.e. things that go in the direction opposite to your main thesis, be careful to put it into only a few well-defined places, reorganizing your argument to achieve that when necessary. Jumping back and forth will confuse the reader unnecessarily. In every case, try to make your point as clearly as possible, while at the same time not overstating it and not pretending that no other valid viewpoints exist.
Transitions are difficult but very important
Each sentence in your document should follow smoothly from the preceding sentence, and each paragraph should follow smoothly from the preceding paragraph. The world is arguably an unstructured jumble of ideas, but anything that you expect the reader to read from start to finish needs to be a linear progression along one single path. Transition words and phrases are what make it possible for a reader to follow you easily as you explore the various ideas in your paper. Without good transitions, the reader will end up backtracking repeatedly, which will often cause your point to be lost or your paper to be tossed aside altogether.
One clue that your writing needs better transitions is if you find that you can cut and paste paragraphs from one section to another without doing substantial rewriting of how the paragraph begins and ends. If making such rearrangements is easy, then you have not been linking your paragraphs into a coherent narrative that reads well from start to finish.
In practice, making smooth transitions is very difficult. Learning to do it takes a lot of practice at first, and actually making the transitions smooth takes a lot of effort every time you write or revise something. One rule of thumb is that whenever you switch topics, you should try to provide a verbal clue that you are doing so, using transitions like “However, …”, “As a result, …”, “For comparison, “, etc.
If you notice that you have to add these words between most of your sentences, not just the paragraphs, then you are bouncing around too much. In that case, you need to reorganize your document to group related thoughts together, switching topics only when necessary. Once the organization is good, all you can do is read and reread what you write, rewording it until each new item follows easily from those before it.
Write what you mean, mean what you write
Speakers use many informal, colloquial phrases in casual conversation, usually intending to convey meanings other than what the words literally indicate. For instance, we often speak informally of “going the extra mile”, “at the end of the day”, “hard facts”, things being “crystal clear” or “pretty” convincing, someone “sticking to” a topic, readers being “turned off”, something “really” being the case, etc. Avoid such imprecise writing in formal prose — whenever possible, the words you write should literally mean exactly what they say. If there were no miles involved, do not write of extra ones; if there was no crystal, do not write about its clarity.
Among other benefits, avoiding such informal language will ensure that your meaning is obvious even to those who have not learned the currently popular idioms, such as those for whom English is a second language and those who might read your writing years from now or in another part of the world. Formal writing should be clear to as many people as possible, and its meaning should not depend on the whims of your local dialect of English. It is a permanent and public record of your ideas, and should mean precisely what you have written.
Unfortunately, specifying minimum page requirements encourages redundancy, but please try to avoid that temptation. When two words will do, there is no need to use twenty. Whenever you finish a sentence or paragraph, read over it to see if any words or sentences can be eliminated — often your point will get much stronger when you do so. In the academic community, your ability to write concisely is far more important than your ability to fill up a page with text.
Academic courses specify page minimums to ensure that you write an essay of the appropriate depth, not to test whether you can say the same thing a dozen different ways just to fill up space. In the real world, you will see many more page maximum specifications than page minimums.
Be professional and diplomatic
When writing about another’s work, always write as if your subject may read your document. Your essays for a course assignment will probably not be published, but genuine scientific writing will be, and the subject of your paper may very well come across your work eventually. Thus it is crucial to avoid pejorative, insulting, and offensive terms like “attempt to”, “a waste of time”, “pointless”, etc.
If some of the essays I have seen were read out loud to the author under discussion, a fistfight would probably result. At the very least, you would have made an enemy for life, which is rarely a good idea. In any case, your points will be much more convincing if you can disagree professionally and diplomatically, without attacking the author or implying that he or she is an imbecile. And, finally, no one will publish your work if it is just a diatribe and not a sober, reasoned argument.
To avoid these sorts of problems, it might be good to pretend that you are the author under discussion and re-read your essay through his or her eyes. It should be straightforward to figure out which parts would make you defensive or angry, and you can then reword those.
Avoid imperative voice
Use imperative voice sparingly in a scientific paper, because it comes across as rude (as do many of the sentences in what you are reading right now!). E.g. do not say “Recall that …”. Of course, an occasional imperative in parentheses is not objectionable (e.g. “(see Walker 1996 for more details).”).
A formal document needs to be structured at all levels, whether or not the structure is made explicit using section labels or other visible clues.
The standard format for an effective essay or article is to: (1) present a coherent thesis in the introduction, (2) try your hardest to convince the reader of your thesis in the body of the paper, and (3) restate the thesis in the conclusion so that the reader remains quite sure what your thesis is, and so that the reader can decide whether he or she was convinced.
Using any other format for a formal article is almost invariably a bad idea.
The introduction and conclusions do not always need to be labelled as such, but they need to be there. Note that an abstract is no substitute for an introduction; abstracts act as an independent miniature version of the article, not part of the introduction.
Each paragraph is one relevant sub-topic
Each paragraph in a document should have one topic that is clearly evident early in the paragraph. Every paragraph should have a clear relationship to the main topic of your document; if not, either the paragraph should be eliminated, or the main topic should be revised.
Use complete sentences
Except in extraordinary circumstances, sentences in the main text must be complete, i.e., they must have a subject and a verb, so that they express an entire thought, not just a fragment or the beginning of a thought. Note that most “-ing” words are not verbs. “The light turning green” is just a fragment, i.e., a start to a sentence or a part of one. To be a sentence that you could use on its own followed by a period, it would have to be “The light turned green”, which has both a subject and a verb.
Put appropriate punctuation between sentences
Two complete sentences can be divided with a period, question mark, or exclamation point, or they can be weakly connected as clauses with a semicolon. However, they can never be connected with a comma in formal writing! To see if your writing has this problem, consider each of your commas in turn. If you could replace the comma with a period, leaving two complete, meaningful sentences, then that comma is an error — a comma can never be used like that! Instead, replace the comma with a semicolon, in case you have two sentences that need to be linked in some generic way, or make the linkage explicit with a conjunction, or simply use a period, to leave two complete and independent sentences.
Section titles for an article should say exactly and succinctly what the reader will get out of that section. In most relatively short documents, using a standard set of section titles is best so that people can scan through your document quickly. Section standards vary in different fields, but a common set is: Introduction, Background, Methods (for an experimental paper) or Architecture (for a modelling paper), Discussion, Future Work (often merged with Discussion), and Conclusion.
If you do not use the standard titles, e.g. if you have labelled lower-level subsections, you should be quite explicit about what is in that section. Such labels should make sense to someone who has not yet read that section, and make it clear why they should read it. For instance, a section about adding a second eye to a simulation of single-eye vision could truthfully be called “Multiple eyes”, but that title is meaningless to someone scanning the document. Instead, it should be something like “Extending the model to explain stereo vision” whose meaning will be clear to the type of person likely to be reading the paper.
Everything important goes in your introduction and conclusion
Everyone who looks at your paper will at least skim the introduction and conclusion, and those who read it in depth will remember those two sections the best. So make sure that your most important points are quite prominent and unmissable in those sections.
Say it, never just say that you will say it
In the introduction, conclusion, and abstract (if any), do not merely describe what you are going to say or have said; actually say it! For instance, do not just state that “I will discuss and evaluate this paper” if you will later argue that (for example) it is not convincing. Instead state that the paper is unconvincing, and (in brief) why you believe that to be the case. Then you can elaborate on that point in subsequent paragraphs.
If you have sections 1, 1.1, and 1.2, there must be an introductory material between 1 and 1.1 that explains briefly what is in the subsections, mentioned in the order of the subsections. That is, 1.1 should never follow just after 1 without some intervening text. If you have 1.1, there must always be a 1.2; otherwise 1 and 1.1 should be merged. Each 1.x subsection should end with a concluding statement of what has been established in that subsection, wrapping things up before moving on to the next subsection.
Different communities have different expectations on what to put into figure captions. Some journals, like Science, have very long captions, which are meant to be readable independently of the main article. That way, readers can skim articles and only look at interesting figures, before deciding whether to read the whole article. In such cases, you must ensure that all of the main points of the figure are also mentioned in the text of the article, so that someone reading the article straight through will not miss them. Other journals and other publications like books, theses, and proposals tend to have very little in the caption, with the figures being understandable only when reading the main text. Even in such cases, I myself prefer to put all the graphical details like “the dotted line represents” in the caption, plus enough context so that the import of the figure is clear. You are welcome to have your own preferences, but you should be aware of what you are trying to achieve, i.e. whether you want the caption to be readable on its own.