Filler words and what to do about them

The instruction to remove every "um" from your speech is delivered constantly and is based on a misunderstanding of what fillers do.

What research actually finds

The picture is more interesting than the advice.

Fillers occur in every language and in all speakers, at rates of roughly a few per hundred words in ordinary speech.

They tend to appear before complex or less predictable words, which suggests they are a by-product of planning speech in real time.

Some studies find that listeners actually recall following content slightly better when a filler precedes it, apparently because it signals that something is coming.

And listeners generally do not consciously notice fillers at typical rates — they notice them when the rate is high or when they are listening for them.

Which means the objective is reduction to an unremarkable level rather than elimination.

When they do become a problem

The genuine cases.

Frequency high enough to be noticeable, which is roughly where a listener starts counting.

A single filler used repetitively, which is more noticeable than variety.

Fillers at the start of every sentence, which reads as hesitancy.

Fillers in recorded material, where they are more prominent than in live speech.

And formal contexts where the expectation is more polished.

The underlying causes

Which determine the fix.

Speaking faster than you are thinking, which is the most common cause by a distance.

Discomfort with silence, so the gap is filled.

Insufficient preparation, so the structure is being invented live.

Nerves.

And habit, which persists after the original cause has gone.

Each responds to a different intervention, which is why generic advice fails.

The pause

The actual replacement.

Fillers occupy the space where a pause belongs.

Which means the fix is not silence-with-effort but learning that a pause is acceptable and is in fact better.

Pauses give the listener processing time, create emphasis, and read as confidence.

They feel far longer to the speaker than to the listener — a two-second pause feels enormous and is barely noticed.

And practising deliberate pauses is more effective than practising the absence of fillers.

The practical method

In order.

Record yourself and count, since almost everyone is wrong about their own rate.

Slow down, which alone removes a substantial proportion.

Prepare more, particularly the transitions between sections, which is where fillers cluster.

Practise pausing deliberately at sentence ends.

Get feedback — someone raising a hand each time, in practice sessions, which produces awareness rapidly.

And re-record after a few weeks, which is the only reliable measure.

What not to do

Counterproductive approaches.

Monitoring yourself while speaking, which splits attention and generally makes both the fillers and the content worse.

Replacing "um" with "so" or "right", which is the same behaviour with a different sound and is frequently more irritating.

Speaking faster to avoid the gaps, which produces more fillers.

And anxiety about it, which is itself a cause.

Other verbal habits

Which matter more than fillers.

Upspeak — rising intonation on statements — which is read as uncertainty in some contexts and is a normal feature of some dialects.

Repeated qualifiers: "sort of", "kind of", "I think maybe".

Apologetic openings: "sorry, this might be a stupid question".

Trailing off rather than finishing sentences.

And these undermine authority more than fillers do, while receiving far less attention.

The apology habit specifically

Worth separating out.

Prefacing contributions with apology or self-deprecation reduces how the contribution is received, and it is a habit rather than a personality.

The fix is direct: state the point without the preamble.

"I think we should reconsider the timeline" rather than "sorry, this is probably obvious, but I was just wondering whether maybe".

Which is the single change that most alters how someone is heard in meetings.

Perspective

Worth keeping.

Listeners are considerably more tolerant of fillers than speakers assume, and considerably less tolerant of speech that is too fast, monotone or unclear.

Which means anyone with limited practice time should spend it on pace and variety rather than on eliminating "um".

And a talk with fillers and genuine engagement beats a fluent one delivered flat, every time.