ACM Interactions

I’ve only posted twice to the Interactions blog in 8 months, but I’ve been writing, and frequently thought “this would be a good blog essay.” Minutes ago, I emailed in the last proof edit for a book. This post covers things I learned about writing and the English language after a brief, relevant description of the book.

From Tool to Partner: The Evolution of HCI is being published by Morgan Claypool. Twenty-five years ago I agreed to update the “intellectual history” that Ron Baecker wrote for the first edition of Readings in HCI. He had cited work in different fields; the connections and failures to connect among those fields mystified me. I didn’t have enough time to resolve the mysteries, and other questions surfaced later. I tracked down people to interview and eventually answered, to my satisfaction, each of my questions. When I was invited in 2002 to write an HCI encyclopedia article on social software, I asked to write about HCI history instead. An article in IEEE Annals of the History of Computing and several handbook chapters followed. In early 2016, I decided to update and extend that work, looking across HCI as it is practiced in human factors, management information systems, library and information science, and computer science, with a tip of the hat to design and communication. What do these fields have in common? What has often kept them apart? How did each evolve? A good way to learn about your own field is to understand how it resembles and differs from others.

Surprising things I learned about the English language while working on this book and the handbook chapters are general insights, but they came into focus because writing about history is different from other writing. Here too, contrast brought clarity.

An author may interact with an editor, copy editor, proofreader, compositor, and external reviewers. The focus on language is greatest with copy editors, who clean up punctuation, word choice, grammar, sentence structure, citations, and references. Some handbook publishers outsource copy editing to fluent but not necessarily native English speakers who contract to apply The Chicago Manual of Style. This book’s editors were native speakers who did an excellent job, yet issues arose. Because the editors were good, I realized that inherent ambiguities in English exist and may be irresolvable, although some could be addressed by applying more sophisticated rules.

Years ago, I broached this with someone at The Chicago Manual of Style, who objected strongly to publishers mandating adherence to their guidelines. The guidelines do not always apply, I was told. “The Chicago Manual of Style itself does not always follow The Chicago Manual of Style!” Nevertheless, copy editors need a reliable process, and if adhering to the extensive CMS is insufficient, what can be done? Authors: Work with a copy editor sympathetically but be ready to push back in a friendly fashion. Copy editors: Let authors know you are applying guidelines and consider additional processes. 

Context matters

We know that goals differ among readers of blogs, professional and mass media magazines and websites, conference proceedings, journal articles, and books, though some lines are blurring. The following examples show that even within one venue, a book, goals can differ in significant ways.

Writing about history for scientists is different from scientific writing to inform colleagues. The treatment of citations and dates are two strong examples. Science aims for factual objectivity, whereas a history writer selects what to include, emphasize, and omit. In scientific writing, the number of supporting references and their authors’ identities can be highly significant. Engaged readers may track down specific references; they may frequently pause and think while reading. The reader of a history (unless it is another historian) is looking for the flow of events. Lists of citations interrupt the flow, whether they are (name, year) or just [N] style. Rarely do readers of a history consult the references. To smooth the flow, histories often move secondary material as well as citations to footnotes or endnotes, whereas scientific writing usually keeps material in the text. Popular histories, including some by outstanding historians, now go so far as to omit citations, footnotes, and endnotes from the text while collecting them in appendices with sections such as “Notes for pages 17-25.”

Dates take on greater significance in a history. When something happened can be more important than what happened. Often, I wrote something like “In 1992, such and such was written by Lee,” and a copy editor changed it to “Such and such was written by Lee (1992). The details of such and such were not important, only that Lee worked on it way back in 1992. The emphasis shifted, the key point is buried. Similarly, in discussing a period of time, a work written in that period is very different from a work written later about the period. One is evidence of what was happening, the other points to a description of what was happening. The date is crucial for the first, not for the second, which could be relegated to an endnote.

History contrasted with science is a special case, but subtle distinctions of the same nature could affect established vs. emerging topics, or slow-moving versus rapidly-evolving fields. In a dynamic research area, the year or even the month a study was conducted can be important, a fact that we often overlook by not reporting when data were collected.

The strange case of acronyms

I will work through the most puzzling example. We all deal with acronyms and blends (technically called portmanteaus, such as FORTRAN from FORmula TRANslation). HCI history is rife with universities, government agencies, professional organizations, and applications that have associated acronyms. A rule of copy editing is that the first time such a compound noun is encountered, the expansion followed by the acronym appears, such as National Science Foundation (NSF) or Organizational Information Systems (OIS). After that, only the acronym appears. Copy editors apply this global replace. Unfortunately, in many circumstances it is not ideal or even acceptable. Some are context-dependent, so a copy editor can’t know what is best. In some situations, almost everyone in the author’s field would agree; in others, perhaps not.

One important contextual issue is the familiarity of the acronym to typical readers. Many systems and applications were known only by their acronym. Discussions of Engelbart’s famous NLS system rarely mention that it was derived from oNLine System, or that IBM’s successful JOSS programming language stood for JOHNNIAC Open Shop System. To introduce the expansion is disruptive and unnecessary. I was familiar with the pioneering PLATO system developed at UIUC (aka University of Illinois at Urbana-Champaign : ), but did not know the expanded acronym until last week, when I looked it up for an acronym glossary that reviewers requested I include with my book. The rule of including an expansion only once is fine for a short paper or for an acronym that is familiar or heavily used, such as NSF and CHI in the case of my book. But when a relatively unfamiliar acronym such as OIS is introduced early and reappears after a gap of 50 pages, why not remind the reader (or inform a person browsing) of the expansion? Override the rule! Similarly, an acronym that is heavily used in one field, such as IS for Information Systems, could be expanded a few times when it might be confused by people in other fields (e.g., for information science). These judgment calls require knowledge that a copy editor won’t have.

Other examples are numerous and sometimes subtle. CS is a fine acronym for Computer Science, and CS department is fine, but “department of CS” is awkward and “the Stanford Department of CS” is even worse. Similarly, it seems shabby to say that someone was elected “president of the HFS” rather than “president of the Human Factors Society.” We would not call Obama “President of the US” in a formal essay.

I will conclude with one that I haven’t been able to solve. Consider the sentence, “The University of California, Los Angeles won the game.” That sounds good. But “The UCLA won the game” sounds worse than “UCLA won the game.” Global acronym replacement fails.

Try this experiment: In the sentence “X awarded me a grant,” replace X with each of the following: Defense Advanced Research Projects Agency, Federal Bureau of Investigation, Internal Revenue Service, Indiana University, University of California Los Angeles, National Science Foundation, DARPA, FBI, IRS, IU, UCLA, and NSF. Which sentences should start with the word “The”? I would do so for all of the expanded versions except Indiana University. For the acronyms, I would only do it for FBI and IRS. Global replace often fails, and I cannot find an algorithm that explains all of these. My conclusion is that English is more mysterious than I realized, and authors are well-advised to pay close attention and collaborate sympathetically with editors.

Jonathan Grudin