January 31, 2018
Let's change our attitude about the Word grammar checker
A fairly regular occurrence in the editor groups I participate in on Facebook or Twitter is that someone posts about how the spelling or grammar tools in Microsoft Word have gotten something spectacularly wrong. As I've noted before, editors in particular seem to take glee in bashing the grammar checker.
I find this frustrating for a couple of reasons, and I've pushed back a bit on social media when I see posts that dismiss proofing tools. But I thought I owed it to people to explain why I think it's not productive exercise to bash the tools. (Modulo the entertainment value of hilariously bad advice, which is by no means limited to advice dispensed by tools.)
I'm going to focus on two issues: the measurable deficiencies of the tool, and the question of audience.
Measureable, reproducible deficiencies
Imagine that you are a program manager (PM) at Microsoft whose job it is to improve the grammar-checking tool. It's not news to you that the tool gets things wrong sometimes.
You go out into the world to find out what sorts of problems people are having with the grammar checker. You find no shortage of complaints. An article in Slate claims that the grammar checker "makes your writing worse." At least that article has some explicit examples. Other complaints are more abstract:
Put yourself in the place of that PM. How is the grammar checker "creating problems"? Which things is it flagging that are "not errors"? What are examples of the "incorrect options as solutions"? These types of generalized, "it just doesn't work!" observations don't help the PM learn anything specific about what to fix. And they don't necessarily help other users, either, since it doesn't tell a user what to look out for.
As a PM, I might also have some questions about how prevalent these errors are. What percentage of the time does the grammar checker correctly flag errors? 15 percent of the time? 40 percent? 85 percent? In other words, are the issues that people report exceptions in an otherwise functioning tool, or are errors the norm? (If errors are the norm, I as PM would be surprised, since it's not like the company doesn't test the grammar checker.)
There's also a question about what constitutes an error. The grammar checker can find actual, non-controversial errors, like subject-verb disagreement. It can also check style—things like the use of passive:
As the PM, I might ask people to adjust the various knobs and levers to see whether the "mistakes" made by grammar checker are simply suggestions that they disagree with.
If I were that PM, I'd say that sure, please let us know where you're running into issues with the grammar checker. But be specific. Show us the error. Will we be able to reproduce it on the current version of Word? If you adjust the settings, do you still see the issue? And I would ask that in addition to posting on social media about the error, why not engage with the Microsoft community to see if your issue is known? In short, I as PM would ask you to use your experience to help make the product better.
Let's move on and talk about who the audience is for the grammar checker in Word. Let me posit this: the grammar-checker tool in Word is not designed for professional editors. No, let me take a step back: Word itself is designed for corporate use. When the Office team thinks about improving their product, they're not thinking primarily about freelance editors, or novelists, or programmers who are creating README files. They think a great deal about employees at companies that are going to buy 500 or 5000 or 50,000 site license of Office. Merge tools for mailing labels? Two-click tables of contents? Let's face it: the prototypical Word user is someone who sits in a cubicle.
My point is that I think the prototypical user is not expert in spelling or grammar. Nor are they producing works of art— no, they're working on reports and memos and other contributions to the great stream of prose that moves corporations forward every day.
I note this because one editor I know said that she had once taken an article by Louis Menand (Harvard faculty, writer for the New Yorker) and run it through the Word grammar checker. She reports that the grammar checker had substantially worsened the article.
This does not surprise me. Let me give you an analogy. Suppose you want to make a dress. You buy a dress pattern and carefully follow all the directions. When you're done, you have a dress! But you do not have a Pierre Cardin dress, or a Dior, or a Vera Wang.
A dress pattern lets a person of modest skills produce a functional finished product. A dress pattern in the hands of an ordinary person does not produce a work of art. Similarly, the grammar checker in Word helps a writer of ordinary skills produce workable copy. It is not designed to help an ordinary writer produce extraordinary prose. It will not turn the average denizen of a corporate cuberhood into Louis Menand.
We professional editors cannot make broad judgments about a tool because we think it isn't as smart as we are about grammar and style. (Though it is probably more thorough.) We have to consider whether it's useful for its intended audience, and gauge the tool in terms of how well it helps that audience.
Finally, we have to help people use the tool to its best advantage. If someone who is unskilled is using some tool incorrectly, you don't say to them "Yeah, that tool sucks." You teach them how to use it right. How about if we do that?