The -ly trap: how to find and fix adverb overload

"The road to hell is paved with adverbs," wrote Stephen King, and a generation of writing advice has repeated it ever since. The instinct is sound but the rule is too blunt to act on. Not every adverb is a sin, and "delete all adverbs" produces prose as lifeless as the purple writing it's meant to cure. The useful version of the advice is narrower: find the adverbs that are propping up weak verbs, and replace the pair with one strong verb. To do that, you first have to find them.

Why -ly adverbs draw fire

Most problem adverbs share a tell: they end in "-ly" and modify a verb that should have carried the meaning on its own. "He ran quickly" is weaker than "He sprinted." "She said softly" is weaker than "She whispered." The adverb isn't adding information; it's patching a verb that was the wrong choice. Each "-ly" word is a small signal that a more precise verb was available and skipped.

"Walked slowly" → "ambled" or "trudged". "Looked angrily" → "glared". "Closed the door loudly" → "slammed the door". In every case the rewrite is shorter, more vivid, and shows rather than tells. That's the edit worth making, and an adverb count is how you find the candidates.

What an automated detector actually flags

The naive approach — flag every word ending in "-ly" — is wrong, because English is full of "-ly" words that aren't adverbs at all. "Family", "lovely", "supply", "rely", "jelly", "ally", "bully" all end in "-ly" and none of them is an adverb. A detector that counted them would inflate your rate and send you chasing edits that don't exist.

So our adverb detector flags every "-ly" word but subtracts a stoplist of common non-adverbs — the "family / lovely / jelly" set — before reporting. What's left is a much cleaner approximation of your true adverb load. It then computes an adverb rate: the percentage of words in your text that are adverbs. A single number you can track from draft to draft.

Strict mode and the false-positive trade-off

Even after the stoplist, edge cases slip through. To tighten things further, the detector offers a strict mode that adds a check: does the word have a recognisable adjective stem? Real "-ly" adverbs are usually built by adding "-ly" to an adjective — "quick" → "quickly", "soft" → "softly". Requiring that stem cuts false positives sharply.

The trade-off is that strict mode can miss neologisms and rarer formations whose stems it doesn't recognise. So the choice is the usual one: loose mode catches more and over-flags slightly; strict mode is cleaner but may undercount. For a quick edit pass, loose mode is fine — you're eyeballing each flag anyway. For a defensible rate you want to report or compare, strict mode is the safer number.

What rate should you target?

Adverb rate is most useful as a relative measure — is this draft heavier than my usual? — but there are rough benchmarks:

• Under 4% — lean, Hemingway-style prose. Most "-ly" words here are earning their place.
• Tight ad copy often runs under 2%, because every word is fighting for space.
• Above 5–6% — worth a closer look. You're probably leaning on adverbs to do work your verbs should be doing.

These aren't laws. Dialogue, certain voices, and some genres carry more adverbs naturally. The number is a flag for review, not a verdict.

When an adverb earns its place

Keep the adverb when it genuinely changes the meaning rather than amplifying it. "He spoke quietly" is fine if there is no single verb that means "spoke quietly" with the right register — "whispered" and "murmured" carry connotations you may not want. Keep it when removing it loses information: "She almost finished" can't lose "almost", and "He only asked once" needs "only" to stand. And keep it in dialogue and casual voice where mechanical de-adverbing would sound stilted and robotic. The goal is leaner verbs, not zero adverbs, and the detector exists to help you make that call faster — not to make it for you.

A practical editing workflow

1. Paste your draft into the adverb detector and note the rate. That's your baseline.
2. Use the tool to find the sentences carrying the heaviest adverb load — those are the densest targets and the fastest wins.
3. For each flagged adverb, ask: is there one verb that means "verb + adverb"? If yes, swap them. If no, leave it.
4. Re-run and watch the rate fall. A heavy first draft can usually drop a percentage point or two in a single pass without any loss of meaning.
5. Stop when the remaining adverbs are the ones doing real work. The number is a guide; your ear is the judge.

For a fuller editing toolkit, pair the adverb pass with a passive voice detector — passives and adverbs are the two most common ways weak verbs hide — and check the overall result against a readability score to confirm the leaner prose actually reads easier. Tighter verbs almost always do.