Writing software is a pretty mixed bag. I use Grammarly largely because it’s convenient, and usually pretty good at spotting typos and other clangers. I also usually disagree with about half of its grammar and usage issues.
When it comes to style, I take issue with just about everything. I hit “Dismiss” a lot. There are minor glitches in core language usage, but nothing too bothersome. It doesn’t quite get contracted usage in sentences, like “Policy is to… whatever”. It’ll say, “The policy…” which is a bit antiquated.
Grammarly isn’t entirely responsible for this. These are traditional benchmark grammar issues. The trouble is that these benchmarks are way off these days. Writing styles have become mired in old 1990s usage theory, which is turning into a form of semantic sclerosis.
There’s a wonderfully squalid history of Grammar Edicts from The Grave to explain my perspective:
About 1991-ish, early style guides started getting ultra-pedantic, and a bit illiterate. I sat through several of these droning sessions in various forms of training. They were all the same and repeatedly used the same expressions.
For example; the following are almost-verbatim quotes from back then:
“What does “best practice” mean? …We’re not experts! It’s a useless expression…”
No, it isn’t, and some of us are experts. “Best practice” refers to a virtual sea of integrated B2B and professional references to system-level best quality controls, including compliance, audit, checks and balances in business or any methodology. It’s two words to cover the basis of whatever operational practices to which you refer. It’s best practice to refer to best practice.
“(Any subject) …We’re not mind readers!…”
No, you’re just lazy. You’re supposed to be communicating using a language. Interpretation is the basis of language. To that extent, you and your readers are trying to read minds by reading meanings as two-dimensionally as possible. Good luck with that.
“You don’t need words like really, literally, actually, very, etc.”
Wrong. These words are either emphases, nuance or referring to specific cases. “They really did that” is specific to an event or incident. “They literally do not think of these things” is nuance. Either way, these expressions are qualifiers.
I don’t hold it against Grammarly or any other software for getting stuck with these anachronisms. That’s what style guides do; they embalm usage.
The oldest style guide is AP style, from memory. This was originally an innocent attempt to create a quality standard for AP media. Fair enough. It’s a series of editorial benchmarks and formatting, and it’s very reliable.
Since then, a plague of utterly useless style guides has arisen. They’re pretty much the same and based on similar principles. They’re also highly idiosyncratic, trying to be different, and failing dismally.
For instance – They cut out commas on a routine basis. They’ve never heard of semicolons, which are actual pauses in statements. If you tried to read this “revised” stuff aloud, you’d be doing it in a monotone and auto prompt speeds. That’s just getting lost in the pedantry.
Non-scholarly text is in theory supposed to be written like normal speech. There are variables in delivery times, emphases, terminology, and often in styles of delivery to an audience.
Another issue is clarity. This is an each-way bet. On average, I might use 1 in 4 of the clarity suggestions, simply because they are clearer or more efficient. I draw the line between simply rearranging words and losing the style of expression.
To be strictly fair to the writing software:
The general level of literacy and language usage is so low that things like Grammarly are highly effective at the de-disasterization of text. Even when I often disagree with every single issue raised, it really does work well to that extent.
That said:
- The 1990s were 30 years ago. A lot of new usage has happened since.
- Language is and must be fluid and able to adapt to new contexts.
- AI writing is always vulnerable to new usage.
- Software must keep up with the evolution of language.
- Pedantry isn’t particularly useful in any language.
- Any degree of obsolescence reflects negatively on the software products.
Solutions:
“Ongoing learning” for large language models. They’ll need it.
Lose the style guides and focus on expression.
Abbreviations and contracted sentences should be factored in.
Just lose the prehistoric pedants.
_________________________________________________________________
Disclaimer
The opinions expressed in this Op-Ed are those of the author. They do not purport to reflect the opinions or views of the Digital Journal or its members.
