paving the way for

How to spot it

Read the sentence up to the comma, stop there, and ask what somebody would have had to see for the rest of it to be true. That is the test, and it costs about a second. The thing you are testing is a short clause bolted onto the end of a sentence on an "-ing" verb, carrying no subject of its own: paving the way for a merger, making it the cheapest fare on the route, ensuring that the ward stayed open through the winter. The main clause says who did what. The tail says what that led to, without naming one person who watched it land.

Plenty of tails answer the question on the spot. A librarian moves story hour to Saturday and attendance doubles; you can picture the counted heads, and the writer could produce them if you asked. Other tails have no answer at all. When a sentence closes by telling you that a decision cemented a reputation, reshaped a debate, or was paving the way for a new era, there is usually nothing underneath to look at, because nobody has ever watched a reputation get cemented.

So finish the test by promoting the tail back into a sentence of its own, with a subject and a tense. Who paved what way, and in which year? A tail that survives the promotion was carrying a fact all along and merely travelling light. A tail that turns into a sentence you would be embarrassed to publish was carrying a mood.

There is a quieter signal, and it lives in the verb. A writer who actually saw the result reaches for a verb you could film, like doubled or reopened. A writer who did not reaches for a verb that grades instead of reports.

Why models overdo it

Whether a result is worth reporting never enters the choice when a model picks its next few words. The model is weighing which continuation is likely, and after a sentence that states a plain fact, a clause announcing the fact's importance is very likely indeed. The prose it learned from is full of them. Annual reports, ministry statements, funding announcements and press material all treat a bare fact as unfinished business, so the slot after that comma is the cheapest real estate in English. Filling it costs nothing and looks like thought.

None of that is a grammatical fault, and it is worth saying outright before the criticism goes any further. The construction is ordinary English and has been for centuries. What a model cannot supply is the second fact. A writer who wants to claim a consequence has to own one, or go out and find one, and finding one costs an afternoon in an archive or a phone call to somebody who kept the numbers. So the verb does the work the evidence should have done: cementing, solidifying, underscoring, positioning. The sentence ends up with the shape of an argument and none of its contents, and no reader can tell the difference at reading speed, which is probably why the habit survives.

When it's fine

Some results can be checked, and a tail reporting one of those is doing honest work. Barack Obama, at Cairo University in 2009, said that Islam at places like Al-Azhar carried the light of learning through so many centuries, paving the way for Europe's Renaissance and Enlightenment. That is an enormous claim, and the tail does not disguise its size. Promote it to a sentence and you get a thesis historians have argued about for a hundred years, which is exactly the sort of claim a tail may carry, because the argument exists and a reader can go and join it.

James Truslow Adams does the same thing on a smaller scale, ending a long sentence about merchants and country gentlemen learning to work together by calling the habit the thing that paved the way for the Puritan party. He had spent a chapter earning it. Hawthorne shows the other honest use, which claims no significance whatever: Mother Rigby rubs brass on the scarecrow's forehead, thus making it yellower than ever. Nothing there is being graded. It is simply the next thing that happened, delivered in the same breath as the last thing, which is what the construction is for.

A tail with a named witness behind it can stay. When the verb is the only thing holding the claim up, write the second sentence instead.

Where it comes from

Grammarians have a name for this, and it is a grammarian's name rather than an orator's: a free or supplementary participial clause, attached to a whole preceding clause instead of to any single noun inside it. Latin had the tidier version, the ablative absolute, which let a Roman writer fold a completed circumstance into two words and hang it wherever a sentence would take it. English shed the case endings and kept the habit, which is why the participle now floats at the end with nothing agreeing with it. The rhetoric handbooks never issued it a Greek name, because it is not a figure anybody performs. It is a joint.

The joint got busy as English sentences got long and periodic. Hawthorne used it in 1852 to keep a scene moving without starting a new sentence, and Adams used it in 1921 to let a historical judgement ride out of a paragraph of evidence rather than arriving alone. Then the institutional prose of the last century found that the same joint would carry an assertion of importance exactly as easily as a fact, and would carry it past a reader unexamined. That is the inheritance a model trained on a hundred years of published English picks up, and it picks it up without the archive the older writers stood on.