English Fillers: How to Hesitate and Still Sound Fluent

Why silence sounds worse than a filler

The word will not come. You stop. Three seconds pass. The other person waits, then fills the gap with a new question, and the sentence you were building is gone.

A filler does something specific in that moment: it tells the listener that the pause is yours and that speech is still coming. Herbert Clark and Jean Fox Tree argued in a 2002 paper in Cognition that English has a pair of words for exactly this. Uh announces a short delay; um announces a longer one. They are not noise that leaks out when the system stalls, because the evidence in that paper is that speakers place them deliberately, sometimes attaching them to the previous word and sometimes stretching them out.

That is the first thing to unlearn: a hesitation sound is not proof that your English is weak. Native speakers use them constantly, and they use them on purpose.

What fillers actually do: three jobs

Most lists of fillers are alphabetical, which makes them useless. Sort them by the job they do and the list becomes short.

Job one: buying time

Well, let me think. That's a good question. Just a second. Hmm.

Plus the two sounds themselves: um and uh. These buy between half a second and two seconds. Under a second, listeners usually do not notice. Past two seconds, the pause starts to look like a stop.

Job two: softening a claim

I mean, it depends. Sort of. More or less. I'd say probably.

Here the filler is doing grammar work: it lowers the confidence of the sentence without changing its content. Useful when you are giving an opinion you are not sure the listener shares, and awkward when the listener wants a number or a yes.

Job three: self-correcting

Sorry, I mean the blue one. Actually, that's not right. What I'm trying to say is this.

These are not hesitations at all; they are repairs, and they make you sound more precise rather than less. A speaker who corrects a word once looks careful. A speaker who corrects every second sentence looks unsure.

The short list: neutral, informal, regional

Filler	Register	Where you hear it	Risk
um, uh	neutral	everywhere	becomes a tic when repeated
well	neutral	everywhere	can sound evasive at the start of an answer
I mean	conversational	UK and US	overused as a sentence starter
you know	informal	UK, US, Australia	irritates listeners when repeated
like	very informal	younger US and UK speakers	reads as filler in formal settings
sort of, kind of	conversational	UK, US	vague in writing
actually	neutral	everywhere	can sound corrective
honestly	conversational	everywhere	claims sincerity you may not need

Two more are regional and worth recognising but not borrowing: eh at the end of a sentence in Canada and New Zealand, and innit in parts of the UK. Used by a learner, both usually sound like an impression.

Rate control: how many fillers a minute is normal

Published estimates put disfluency in conversational English at about six events per hundred words, counting repeats and repairs as well as fillers. In a corpus analysis of ordinary conversation, average speech rate ran at about 194 words a minute for speakers in their twenties and about 169 for speakers in their sixties. Run the arithmetic on the middle of that range and six per hundred words comes out at roughly ten events a minute.

Ten a minute is not a target. It is context: hesitation is normal, and a learner who says nothing rather than risk a filler is solving the wrong problem.

For practice, use a band instead of a rule: no more than two or three deliberate fillers a minute, and never the same one twice in a row. Above that, listeners start counting instead of listening, and hesitation stops being a strategy and becomes a habit. That is how to speak English without hesitation in the sense that matters: not zero fillers, but fillers you chose.

How to speak English fluently free of the um habit

Three substitutions do most of the work.

1. Replace the sound with a short silence. A half-second pause inside a clause is almost inaudible; um draws attention to it.
2. Replace the sound with a phrase that carries meaning: Well, Let me think, The thing is,. These do the same job and add a beat of structure.
3. Reduce how much you have to plan. Say one clause and stop. I went there last year. then It was crowded. Two short sentences need less look-ahead than one long one, and look-ahead is exactly what fillers are covering.

There is a fourth move that sounds counter-intuitive: slow down at the start of the sentence rather than in the middle. Most fillers land just before the hard word, and the hard word is easier to reach if the words before it were not rushed.

Replacing your first-language hesitation sounds

Most learners do not hesitate in English. They hesitate in their first language and then translate the sound. Spanish and Portuguese speakers often produce a long "ehh", French speakers a nasal "euh", German speakers "aeh", Japanese speakers "etto" or "ano", Korean speakers "geu", Mandarin speakers "neige" or "zhege".

The problem is not the sound itself. It is that the sound announces a switch: listeners hear the language change, and the sentence loses its thread. Swap the sound for an English one and the same delay costs you much less.

Two steps. First, record yourself answering a question for one minute and count the sounds that are not English. Second, pick one English replacement, well if you need a beat or um if you need a second, and use only that one for a week. One replacement is easier to automate than five.

A two-minute recorded drill with a filler count

You need a phone, a timer and a question you have not prepared. What did you do last weekend? works; so does What's a skill you'd like to learn?

1. Record two minutes without stopping. Do not edit out the hesitations; they are the data.
2. Count the fillers: um, uh, you know, like, I mean, and any first-language sound.
3. Count the words, then divide by two to get your words per minute.
4. Divide the fillers by two to get fillers per minute.
5. Write both numbers down and repeat the drill three times in a week with a new question each time.

Measure	What it shows	What to do with it
Fillers per minute above four	hesitation is clustering	substitute a short silence for the sound
Fillers per minute under one	you may be avoiding risk	keep a phrase ready, not a silence
Words per minute falling	planning load is too high	shorten the sentences, not the vocabulary

The target band for the drill is one to three fillers a minute with no repeats in a row. When you hit it, change the question to something harder rather than trying to reach zero.

Frequently asked questions

Is it bad to use 'um' and 'uh' when speaking English?

No. They are ordinary English words, used by native speakers in every conversation, and the research treats them as signals rather than errors. What costs you is repetition: the same sound in every sentence, or one that runs for two seconds while you search. In an interview or a speaking test a filler is much cheaper than silence, because silence reads as a lost turn.

What can I say instead of 'um' in English?

For a short delay, well or hmm. For a longer one, Let me think. or That's a good question. While you choose a word, I mean or What I'm trying to say is holds the turn and tells the listener that a repair is coming. If you need a beat and nothing else fits, a half-second silence is the safest replacement, because it leaves no sound for a listener to count.