The forms, levels and principles of editing

Editing comes in many forms whose names are not settled: forms told apart by what they deal with, levels of edit that divide the work among specialists, and passes through the text. Medical editors, in one experienced editor's observation, mostly distinguish only copyediting and substantive editing, and substantive editing can look for errors in the text as a whole, in each paragraph and in each sentence. Medical editing grew up apart from the technical editing of the physical sciences and engineering, whose literature supplies both a vocabulary for these forms and principles an editor can work by, among them where an editor stands on the author's voice.

Medical and technical editing

Medical editing developed separately from the technical editing that supports the physical sciences and engineering, which deals with a wider range of documents, readers, editorial tasks, working environments and educational opportunities. ¹ Most books on technical editing are written for students on technical-writing degrees who expect to work as technical writer-editors in research, manufacturing or technology companies, often in publication departments. ² Many technical editors work with several other specialists to prepare and manage the production of long, multi-authored and complex company-sponsored reports, and most enter the field with some formal training or experience in technical writing or editing. ² Many if not most medical editors, by contrast, are author's editors who work mainly with authors on scientific articles and typically work in clinical or research departments or are self-employed and work from home, so they often have different degrees of editorial freedom and autonomy from technical editors. ² Many if not most medical editors also learn the craft on the job, and so are often unaware of developments and resources in the larger field of technical editing. ¹

Forms and levels of editing

No form of editing has a broadly accepted definition, and 30 books on technical editing use 77 terms for the forms, from accuracy, analytical and author's editing through copy clarification, developmental, heavy, light, line, macro- and micro-, rule-based and substantive to typographic and usage editing. ³ Most of the terms refer to editing content (what is said), presentation (how it is said), correctness (how properly it is said) or formatting (what it looks like). ³ Others classify editing by level of edit, by degree of effort (light, medium and heavy), by stage in the publication process (pre-submittal, author's editing and post-submittal) or by requested turnaround — rush, urgent and "you've got to be kidding". ³

The level of edit divides editing into forms by the kind of content they deal with, which at least in theory need different skills and amounts of time to evaluate, do not unduly overlap and may be completed by different specialists, such as managers, copy editors and production specialists. ³ Industry uses the concept widely to help schedule and budget editing services and set priorities. ³ Most proposed sets of levels draw on the nine introduced by the Jet Propulsion Laboratory in 1976: coordination (project management), policy (company requirements), integrity (cross-references in place), screening (light copyediting), copy clarification (notes to compositors), format (headings and typographic details), mechanical style (consistency), language (full copyediting) and substantive (content and organisation), where the glosses in parentheses are a later reviewer's, not the laboratory's. ³

Editing can also be divided into passes, each a beginning-to-end readthrough of the text in which particular tasks are completed. ³ A more functional approach to passes or levels sorts the issues by when they can be found: turning pages (formatting, page design), skimming (headings, organisation), scanning and reviewing (specific information, consistency), reading (copyediting, cross-checking), analysing (substantive editing, rewriting) and verifying (fact checking, peer review). ³

Copyediting and substantive editing

Most medical editors, in one experienced editor's observation, generally distinguish only copyediting and substantive editing, which need fundamentally different perceptual skills. ³ Copyediting is literally looking at and correcting the ink on the page, and needs a narrow external focus of attention on something that can be seen. ³ Substantive editing is finding and organising ideas on the page, and needs a broad internal focus of attention on logic and abstract concepts processed cognitively. ³

One guide to English editing defines copyediting as checking a text for accuracy and consistency, and for whether its layout conforms to specific guidelines. ⁴ It concerns the visual characteristics of the text, including headings, bullet points, spelling, punctuation, parentheses, references, figures, tables and appendices. ⁴

In one writer's practice, substantive editing, which that writer sometimes calls reconstructive editing, takes a very badly written and organised document and reworks it completely to make it readable. ⁵

Substantive editing at the text, paragraph and sentence level

The guide to English editing that defines copyediting treats substantive editing at three levels: the text as a whole, the paragraph and the sentence. ⁴ Its advice is not limited to one document type, so the problems it covers, in one reviewer's judgement, bear on regulatory and medical communication documents too. ⁴

At the text level, the editor checks for a main topic statement with logic and cohesion followed by a concluding statement. ⁴ Common errors at this level include unfulfilled announcements and missing markers, where the writer announces a topic and then fails to discuss it, and insufficient differentiation of unfamiliar information, where the writer wrongly assumes the reader already knows information not yet introduced. ⁴ Muddled use of singular and plural, mixed positive and negative statements, inconsistent terminology, insufficient repetition and unclear pronouns can all contribute to cumbersome and ambiguous reading. ⁴

At the paragraph level, substantive editing ensures that each paragraph presents a single cohesive idea. ⁴ Writers are particularly prone at this level to dividing paragraphs incorrectly, not defining the paragraph's topic in its first line, misusing transitional statements between several points within one paragraph, and ordering the information within a paragraph incorrectly. ⁴ Parallel terminology, where different words describe the same thing within one paragraph, increases the complexity of the text and can affect its readability. ⁴ Its advice on each of these features is meant to keep the flow of the text logical, coherent and unambiguous. ⁴

Substantive editing at the sentence level is similar to editing at the paragraph level. ⁴ Here the editor checks for features including incorrect use of pronouns, incorrect ordering of information, the correct presentation of parallel ideas, inappropriate connective words, empty wording, passive wording and ambiguity. ⁴

Principles of editing

One reviewer of the technical-editing literature has proposed from it 23 principles of medical-technical editing, with a slight bias toward author's editing and scientific publications. ^{1, 6}

1. Remember the goal: technical communication is functional communication that helps readers accomplish tasks or make decisions, and the goals of editing are to help readers understand, find, remember and use information; the editor must ensure that the information is relevant, accurate, complete and accessible and the writing clear, organised, concise and grammatically correct; the information must be necessary and sufficient for its purpose, which means that ideally a text should be as long as it needs to be, or is allowed to be, but as short as it can be made.
2. Understand the assignment before starting: identify the primary audiences, purposes, deadlines, format, expectations and importance of the text; audience and purpose determine everything in a technical document, and the importance, the author's or client's expectations and the deadlines should guide how much time it gets and its priority in the workload.
3. Work efficiently: an editor can have up to four clients — readers, authors, employers and publishers — whose needs can be difficult to balance, and employers and clients are paying for the editor's time and expertise, and the editor owes it to them to work as efficiently as possible.
4. Consider examining a long and complicated text first: checking its length, headings, topic sentences and lists may indicate the scope of the work, by giving a sense of the topic and organisation and identifying tables that need major revision, figures that were not included and other files with supplemental data. Alternatively, especially for a shorter text, begin by editing, so as to go through the same sense-making process readers do, uninfluenced by an earlier review.
5. Take care of authors: editors exist to help them, so a collaborative relationship with them is highly desirable; be tactful when suggesting changes and complimentary where possible — "sell, don't tell"; comments are usually better phrased as questions than as statements, asking "Did an IRB approve the study?" rather than stating "The study should have been approved by an IRB"; and remember that the authors' names go on the publication, not the editor's.
6. Put readers first: meeting their need for information is one of the defining characteristics of technical editing; establish what they want to know, need to know, already know, don't know and think they know that isn't so, and mediate between the authors' need to own the message and the readers' need for information.
7. Guide and focus attention: the essence of good communication is helping readers attend to the right things without being distracted by the wrong things; use the full range of textual, visual and graphic conventions to direct readers to information and emphasise what is important — type fonts and size, bolding and lists; white space, layout and margins; colour, line weight and plotting symbols in graphs; circles and arrows on images.
8. Assume the worst: treat the text as containing nothing of value until proven otherwise and ignore the reputations of the authors, their institutions and the target journal, because an editor who has to recreate the authors' meaning from a baseline of no meaning, and question their assumptions about the topic and readers, is more likely to find missing information, deeper problems of logic and vague or ambiguous meanings.
9. Edit aggressively, recommending numerous or extensive changes or even restructuring or rewriting where the editor thinks it necessary and querying the authors when unsure of the original meaning or the accuracy of a revision, since changing the meaning without acknowledging it can damage credibility — but "If it ain't broke, don't fix it."
10. Edit everything time permits, doing everything possible to improve the text and advise the authors: for example, verify data against source articles, fact-check, suggest highlights on radiographs, create tables, propose examples, and above all shorten and tighten; skilled editing can reduce many texts by up to 30% while maintaining or improving clarity.
11. Justify every edit, specifically — explain why a sentence is vague rather than saying that it is, and cite references, authorities and style manuals; the more aggressive the edits, the more they may need explaining, and justifying them helps authors accept them, improves the editor's credibility and requires the editor to think critically about the editing.
12. Insist on understanding: editors are rarely content experts, but an understanding sufficient to be useful is often possible, and if the editor can understand the text readers almost certainly will; where part of it is unclear, tell the authors or ask whether it has been understood correctly, do not accept the near-universal claim "My readers will know what I mean", and where possible ask the author to explain.
13. Don't "read over", skimming past text the editor does not understand in the belief that the author is probably correct, or "read in", usually unconsciously filling gaps from the editor's own experience with information readers do not have; both are reasons authors need editors, both are prevented by reading more closely, and the text of a skilled author can sometimes lull an editor into reading rather than editing.
14. Eliminate ambiguity: a text whose readers cannot be sure they have understood it correctly may be of no use to them; meaning and ambiguity often depend on context ("She had a solution" means one thing to a chemist and another to a mathematician) and on plausible alternative readings ("Prostitutes appeal to the Pope"): write so that the text cannot be misunderstood, though, as two technical editors who trained the reviewer cautioned, clarifying the meaning sometimes coincidentally reveals it.
15. Look for what is missing: errors of omission, such as how the sample size was determined, are harder to detect than errors of commission such as misspellings; ask, if a statement is true, what must have been true before it and what must be true after, and check for missing data, references to key claims, cross-references, eligibility criteria, table and figure citations, endpoint definitions and gaps in descriptions, explanations and justifications.
16. Stay visually aware: editors need to be as comfortable communicating with visuals as with words, and besides designing slides and posters should be able to enlarge, crop and label a photomicrograph, draw a flowchart, create a table, and know when a line drawing will work better than a photograph.
17. Check the math, as editors have warned of numerical and arithmetic errors in scientific publications for decades: numerical errors were found in two-thirds of 32 consecutive manuscripts submitted for editing at a major medical centre and in two-thirds of 157 articles from 20 radiology journals, and the most common errors are in rounding, sums, percentages and missing or incorrect numbers.
18. Raise ethical concerns: editors "serve science as well as clients" and should advocate for transparency, accuracy and accountability in their work; the concerns that may need raising with authors or others include obviously biased or even deceptive authorship, including guest and ghost authorship, plagiarism, duplicate publication, copyright violations, visual distortions in graphs and images, and misleading explanations and interpretations.
19. Beware unintentional messages, so that how something is written does not overshadow what is written, which is why Chalkduster typeface and "ain't" are not found in scientific articles: short sentences of too many simple words may seem condescending, many uncommon abbreviations or unfamiliar terms presumptuous, and contractions, exclamation marks and slang inappropriately informal.
20. Don't fixate on wording: most editing is related to language, which makes it easy to stay focused on the wording, but the whole text and its images also need evaluating for clarity and organisation, usefulness to readers, compatibility with the standards of the scientific literature and adherence to the sponsor's house style.
21. Proofread one element at a time, because verifying or correcting several features of a text in the same pass can divide attention and lead to mistakes: separate passes may be needed for consistency of terms, numbers and design features and for cross-references, facts and sequences such as reference numbers, and a checklist of the features to reconcile can help.
22. Guard against the "grammar gremlins" that add, delete or rearrange text invisibly to its editor but visibly to the next reader: only putting the text aside for several hours, best overnight, lets the editor see and correct these changes.
23. Reflect on the work: learning can be defined as a change in knowledge or skill as a result of study, instruction or experience, and much of that study and experience comes from thinking about what you do, reflecting on how and why you do it and answering the questions that arise in the work, and ten years of experience should not be one year of experience ten times.

^{6, 7, 8}

The author's voice

Whether an editor should try to preserve the author's style or voice is perhaps the most controversial question in technical editing. ⁷ The recommendation to preserve it whenever possible appears to be a holdover from literary editing, where the author's voice is integral to the literary value of the text. ⁷ Many editors hold that a multi-authored text benefits from a single consistent style, and other opinions run from editing that should be constrained and inconspicuous to the picture of editors who attack a manuscript as if mugging someone in a dark, deserted alley. ⁷ Yet the defining characteristic of technical writing and editing, for the reviewer who proposed the principles above and the authorities that reviewer cites, is meeting the readers' need for information, not preserving the author's style. ⁷ Many authorities, and that reviewer, hold that "ruthlessness is essential" in technical editing, because it produces the best results. ⁷ Ruthless does not mean tactless. ⁷