Behind the Style: Delivering The Chicago Manual of Style’s 18th Edition in PerfectIt

By Chris Ryder, analytical linguist

I know editors everywhere will be envious when I say that, as PerfectIt’s analytical linguist, I had the honour of working with the team at The Chicago Manual of Style to build support for the 18th edition. However, I’m not writing this article to make you jealous! I want to share the technical details of what we did so that you can understand what the style will find and what it won’t.

Here are some of the biggest additions, revisions, and fine-tuning that we’ve undertaken over the last few months to bring PerfectIt up to date with the release of the 18th edition of CMOS.

A bit about me

This is my first article for the PerfectIt blog. By way of introduction, I have an extensive background in linguistics. Before PerfectIt, I worked mostly in academia. I’m very interested in grammar, new words, and the way words change in meaning over time. I also have a lot of experience with corpus linguistics, which is basically a way of using “big data” to reinforce our understanding of language. Large datasets of language can really help us to spot patterns in how words are being used, and this in turn gives us an idea of the kinds of things that PerfectIt can do to simplify the work of writers and editors.

Getting with The Times

The Chicago Manual of Style 18th Edition has new guidance regarding an initial The at the beginning of titles of newspapers, magazines, and other publications – specifically, that it should be included in the italicized title if it appears on the masthead or cover of the publication. But, with hundreds of these publications floating about in print and online, how can you remember which ones include an initial The in their title?

The answer is: you don’t. CMOS for PerfectIt will now search for over 300 newspaper and magazine titles as part of the Phrases to Avoid/Consider check. These titles consist of around 250 from the United States, plus some of the most popular from the UK, Canada, and Australia.

PerfectIt will search for all these titles including The and advise on how each should be presented. For example, it will search for The New York Times and tell you to check that the initial “the” is included as part of the title (as it is here with the initial capital and italics); on the other hand, it will also search for the Chicago Tribune and tell you to check that the “the” is not included. In both cases, there is also a reminder to adjust the context if necessary – you wouldn’t say “the The New York Times article”, so there are exceptions to be aware of.

The two examples above might be fairly easy to remember, but the value of the approach we took is that it covers so many more. You don’t have to remember that it’s The Fresno Bee, but the Garland Texan, or the Cherokee Phoenix. It doesn’t cover everything. But The Chicago Manual of Style for PerfectIt will do a lot of that work for you.

Going wild

Although wildcards can cause some people to break out into a cold sweat or shake at their very foundations, they can be a brilliant way to streamline a large number of checks that follow a distinctive pattern, and to make them extremely comprehensive too. And – I may as well lay all my (wild)cards on the table – I personally quite enjoy them!

The Chicago Manual of Style 18th Edition now recommends capitalizing official designations of time and time zones, and we could have included that within PerfectIt’s Capitalization of Phrases check. But what about when someone writes a shortened version, such as 9:00 eastern or 15:00 central? Here, the shortened versions should have a capital letter, but we definitely don’t want to be capitalizing every single instance of eastern and central that appears in the text, as most of them won’t be referring to the time zones.

This is where a wildcard is ideal: a clear pattern of usage, with eastern, central, or whatever the word is following a time given in a specific format. For example, to find every instance of the above for Central time, we have added a wildcard that looks like this:

Find:                <([0-9]{1,2}:[0-9]{2}) central

Replace:          \1 Central

That might look complicated, but all it’s essentially saying is:

  1. Find one or two numbers, followed by a colon, followed by two numbers, followed by space and the word central (lowercase c).

  2. Replace it with all the time stuff you had before, followed by a space and the word Central (capital C).

There are also new wildcards to help with using a comma (or not) before things like Jr. or roman numerals as part of a person’s name, and Inc. or Ltd. as part of a company name.

For more on wildcards and how to become friends with them, check out this previous blog.

Unhyphenating the clearly unhyphenated

Do you hyphenate compounds formed by an adverb ending in -ly? In The Chicago Manual of Style 18th Edition the preference is not to hyphenate these (as in the subheading above), not even before a noun (clearly unhyphenated terms).

When the modifier of a noun is more than one word, a hyphen is often needed to make sure the intended meaning is understood. Imagine the headline Small-State Lawmakers to Convene Tomorrow without the hyphen and … well, you get the idea. Many phrases like this are formed where the first word is an adverb ending in the suffix -ly, such as clearly unhyphenated, overly enthusiastic, highly paid, or mildly amusing. Major style guides, including CMOS, argue that the -ly ending removes any ambiguity about what type of word is coming next, so the hyphen isn’t necessary.

As part of the update to CMOS in PerfectIt 6, we wanted to find a way to capture as many hyphenated forms of these as possible and correct them to unhyphenated forms. It was tempting (for me, at least) to try a wildcard – there is, after all, a fairly clear pattern that they all follow. The problem with this is that it would also catch a number of “false positives” – instances that follow the pattern but aren’t what we’re looking for. The most common of these would be where the compound is an adjective and a noun, rather than an adverb and an adjective, such as an early-morning walk. This type should have a hyphen, so a wildcard that caught this phrase as well could do more harm than good. And there are many hyphenated noun phrases that would also be flagged: supply-chain, roly-poly, butterfly-effect, and more.

The alternative we came up with made use of my experience with corpus linguistics and enTenTen21, a corpus of 52 billion words taken from internet texts dating up to the end of 2021. We used a list of the 1,000 most common adjective phrases containing -ly, vetted them to remove any that weren’t what we were looking for, and added the final list to the Hyphenation of Phrases check. CMOS in PerfectIt 6 will now find almost 900 adverb compounds with -ly, from actively-managed to wildly-popular, and replace the hyphen with a space.

Fine-tuning the fine-tuning

It’s often the subtle adjustments in software updates that profoundly enhance user experience. We’ve been carefully looking through all the existing checks in CMOS in PerfectIt and have made a few adjustments to add that chef’s kiss to the quality of life for the user.

We’ve added in some “Except for”, “Except before”, and “Except after” rules to ensure fewer incorrect hits. For example, CMOS 18 recommends using a hyphen in the phrase fast-talk … but only when it’s a verb. Although it’s not possible to specify a particular part of speech (noun, verb, adjective, etc.) in PerfectIt, there are some tricks you can use to get pretty close. With the addition of an “Except after” rule for fast talk (unhyphenated) for the words the, a, my, your, and many other determiners, PerfectIt is now much less likely to recommend adding a hyphen to fast talk when it is a noun phrase and there shouldn’t be a hyphen.

Exceptions can also be very useful in avoiding clashes with other checks in the same style. There is an Accents check to replace creme with crème, but there are also checks to replace creme brulee with crème brûlée, creme fraiche with crème fraîche, and creme de la creme with crème de la crème. Without exceptions, the check for creme on its own will find all the instances of those phrases too, which will be annoying when the same phrase gets flagged twice. So we have added brulee, fraiche, and de la creme (as well as all the accented versions) as “Except before” phrases for creme, meaning each check will be done separately and not repeated.

Sometimes, simplifying context can enhance a check’s scope and effectiveness. CMOS 18 recommends hyphenating the term president-elect as a noun. Previously, PerfectIt's searches included “the president-elect” and “a president-elect” to avoid verb confusion. However, this limited approach missed variations like “our president elect” or “America’s president elect.” By broadening the search to “president elect”, we capture more relevant instances. There will still be some false positives that would have happened anyway, but now we won’t be missing any contexts that we do want to hyphenate.

Finally, we’ve been working through many of the existing terms as they appear in Merriam-Webster Dictionary. Where there is not a preference for other stylistic reasons, CMOS usually follows Merriam-Webster and favours the first-listed term in that resource. However, because usage and preferences often change over time, it’s important to review this every now and again to make sure that everything is as up to date as it should be. For example, the phrase data set used to be preferred to dataset, but through popular usage the reverse is now true. So we’ve swapped around the check in PerfectIt to match!

As before, most of the checks throughout the style include links to the source material at www.chicagomanualofstyle.org, and we have worked to ensure that these are all the updated versions for the 18th edition.

We’re thrilled to release the new version of CMOS in PerfectIt, and look forward to your feedback as you explore its new features!

Previous
Previous

Reignite Your Text with the Latest from The Chicago Manual of Style

Next
Next

Bridging Voices in Academic Publishing: Using AI to Work More Effectively with Multilingual Authors