The following line appeared in a recent article in the Guardian:
Researchers who questioned more than 90,000 adults found “complete” compliance with government safety measures, such as physical distancing and staying at home, had dropped in the past two weeks from an average of 70% of people to less than 60%.
Notice the problem? This is a good example of a ‘garden path’ sentence. It leads readers up the garden path before the syntax takes a sudden turn that forces them to rearrange and reprocess what they’ve just read.
Not every reader will go up this particular path, but many will, since it’s natural and reasonable to assume initially that what was found was complete compliance. But the real object of found is the entire clause “complete” compliance . . . had dropped in the past weeks . . . – producing a near-opposite meaning.
The problem is compounded by the length and complexity of text before we reach that all-important had dropped. There is the stacked noun phrase government safety measures and the parenthetical phrase such as physical distancing and staying at home. It all takes us a long way down the path before the abrupt twist.
The solution is extremely simple: Add that after found. Like so:
Researchers who questioned more than 90,000 adults found that “complete” compliance with government safety measures, such as physical distancing and staying at home, had dropped in the past two weeks from an average of 70% of people to less than 60%.
Including the conjunction that tells the reader immediately that “complete” compliance is not the object of found but the subject of a whole clause. This closes off the garden path and signals readers to wait and see what was found about compliance.
Omitting that in contexts like this is a common practice and often inconsequential: We knew that we were safe. They said that it was okay. Other times, when the complement clause is deferred or more intricate, omitting that causes a momentary blip in comprehension: then it should be retained. Sometimes it positively cries out to be.
At least in this case the line’s sense is ultimately clear, even if delayed and initially obscured. Occasionally, omission of that ransacks meaning altogether, leaving readers to figure it out as best they can from background knowledge or probability.
Journalists routinely omit or delete that in situations like this. It’s not a good habit to make automatic. John E. McIntyre’s new book, Bad Advice (see my tweets here), refers to the ‘weird hostility’ to this use of that and advises those in the trade: ‘When that is there and does no harm, take your hands off the keyboard.’
Strunk and White’s famous advice to ‘omit needless words’, beloved by journalists and other writers, is counterproductive when it encourages a mindset that’s too eager to cut without regard to the consequences. It’s unwise to omit words if you lack the judgement to know which words are truly needless.
Writers and editors: Beware the garden path
Sentence first 
Great post. I completely agree on the importance of “that” for maintaining clarity. Thanks for sharing!
Thank you for reading, Isobel!
Very good and helpful. In editing, I often find my authors have that “weird hostility” to the use of “that.” A few are even somewhat hostile to me mentioning the problem and the possible confusion arising from the frequent lack of “that.” But most are open to change, thank God.
Thank you. It’s unfortunate, but understandable, that people get defensive about things like this. I hope that at least some of them reflect on and accept the point later.
I worked on the Times in London with a senior sub-editor who insisted on the relative “that” on every possible occasion, even when it was unnecessary, eg “He said THAT he was unhappy THAT I took the relative pronouns out of the copy THAT he had given me to edit,” when “He said he was unhappy I took the relative pronouns out of the copy he had given me to edit” is perfectly understandable (though I would probably leave “that” in after “unhappy” …) It was nonsensically festishistic, and meant copy would often be two or three lines longer than it needed to be, so other stuff had to be cut out to make ithe story fit.
‘Fetishistic’ seems a fair description of that condition. It’s striking how often and how easily people turn personal preferences – even peculiar ones – into iron-clad rules. There should always be room for discretion.
My personal preference would be to strike the first “THAT” and leave the other two, but I recognize THAT I’m “nonsensically fetishistic” about it, and THAT, to use an American advertising disclaimer, “your mileage may vary”.
Well, it might have garden-path’d me, except for your trip-wire alert.
I can’t cite exactly, but didn’t Heraclitus do the same thing in the
early period of writing; that is, lead on toward an irritatingly vague
or even wrong conclusion which readers would then have to put
right; or even end with no logical conclusion at all, and force
readers to come up with their best guess at what he intended.
Or it might have been fragmented texts rather than intentional.
Ach. I tried to avoid priming readers unnecessarily, but I guess the post title (and even the act of excerpting) set up the line for a close reading. I don’t know about Heraclitus in this respect.
Many people THAT I know make rude comments about the fact THAT I use the word “that” a lot more often than most. I reply THAT my use of the word “that” makes the object of my sentences clear, and THAT I can’t help it THAT I’m not like most modern-day writers, who wallow in a bog of incomprehensible verbiage rather than state things clearly.
I really loathe it when I read something and find myself saying, “Wait, THAT almost but not quite made sense.”
It’s a small, simple fix for a lot of unclarity.
“Including the relative pronoun tells the reader immediately that “complete” compliance is not the object of found but the subject of a whole clause.”
But “that” is not a relative pronoun in your example, since the clause after “found” is a complement or noun clause, not a relative clause.
Thanks for flagging that slip of the brain. Fixed.
It appears that the tendency to do away with ‘that’ in sentences such as the quoted one, even when its use would substantially enhance readability, is more prevalent in American journalism than in the British journalism.
* Please Ignore ‘the’ before ‘British journalism’.
How prevalent the tendency is in different countries is not something I’ve explored. If you have any data on this, I’d be interested to see it.
[…] I showed how omitting the relative pronoun that can lead readers up the garden path. It’s especially likely in headlines because of their […]
[…] I said of another ill-advised omission of that in 2020, in a corrective to Strunk and White’s peevish parsimony, it’s unwise to omit words if […]
[…] a key event in the decline of ‘the’ and ‘of’, is also identified in excellent articles by Stan Carey and Carol Saller as a factor in the incorrect deletion of ‘that’ by people who edit text. […]