Public-speaking advice often treats every “um” and “uh” as meaningless noise. That is too simple.
Filled pauses can distract an audience when they are frequent or repetitive. They can also carry information about what is happening in a conversation. A 2001 paper by psycholinguist Jean E. Fox Tree offers a useful example, but its result is narrower than the claim that filler words are generally good for comprehension.
What the paper studied
Fox Tree’s paper, published in Memory & Cognition, examined how hearing “um” and “uh” affected listeners’ real-time processing of spontaneous speech. The paper reported studies in English and Dutch and measured listeners’ ability to recognize words in the speech that followed a filled pause.
Source: Jean E. Fox Tree, “Listeners’ Uses of Um and Uh in Speech Comprehension,” Memory & Cognition (2001). Read the abstract and publication record.
The result was not the same for the two fillers:
- Hearing “uh” had a beneficial effect on recognition of an upcoming word.
- Hearing “um” had neither a beneficial nor a detrimental effect in these studies.
Fox Tree interpreted the difference as support for a hypothesis about delay: “uh” may signal a short upcoming delay, while “um” may signal a longer one.
The narrow finding: In these studies, “uh” helped listeners recognize an upcoming word. “Um” did not help or harm recognition. The result concerns a specific speech-processing task, not an overall rating of speaking quality.
What the result does not prove
This paper did not test whether audiences prefer speeches with more fillers. It did not measure whether a speaker sounded confident, prepared, credible, or professional. It also did not show that every “uh” helps in every sentence.
That distinction matters. “Listeners recognized an upcoming word more effectively in this experimental context” is not the same claim as “filler words improve public speaking.”
The study also gives no basis for treating “um” and “uh” as interchangeable. They produced different results, which is part of the paper’s central point.
Why a filled pause can carry information
Spontaneous speech unfolds in real time. A speaker may need a moment to retrieve a word, plan a phrase, or decide what comes next. A filled pause can alert the listener that the speaker has not finished and that a delay is coming.
That signal can be informative without being desirable in unlimited quantities. A yellow traffic light conveys information, but a journey made entirely of yellow lights would still be frustrating.
The practical lesson is not to add fillers deliberately. It is to avoid treating every filled pause as identical verbal waste.
What this means for speakers
Look for patterns, not perfection
A single “uh” before a difficult word is different from a repeated “um” between nearly every phrase. Count where fillers cluster and what the speaker was doing at the time.
Transitions, unfamiliar material, and improvised answers may produce different patterns. That context gives you something useful to practice.
Separate awareness from judgment
A count tells you what happened. It does not tell you why it happened or whether the audience found it distracting.
Use the count as a prompt for review. Listen to the relevant moment when a recording is available, or ask an evaluator whether the pattern affected clarity.
Do not turn zero into the only acceptable score
Reducing a distracting habit can improve delivery. Chasing a perfectly filler-free performance can also make a speaker more self-conscious and less willing to speak naturally.
A more useful goal is control: notice recurring fillers, understand when they appear, and become comfortable replacing the distracting ones with a deliberate pause.
What this means for an Ah-Counter
The Toastmasters Ah-Counter role requires contextual judgment. Words such as “so,” “and,” and “like” can perform normal grammatical work or function as verbal fillers. Even “um” and “uh” should be reported as observations, not as evidence that a speaker failed.
That is why a useful counter remains manual and why its final report should emphasize patterns rather than a winner or a zero score.
The bottom line
Fox Tree’s paper is evidence against one sweeping assumption: “um” and “uh” are not identical sounds with identical effects, and they are not necessarily meaningless.
It is not evidence that fillers always help, that speakers should use more of them, or that frequent fillers never distract an audience.
Count them when awareness is useful. Reduce the patterns that interfere with the message. Keep enough nuance to distinguish a human pause from a habit the speaker wants to change.
Put this into practice
The speaking.app Filler Counter lets an Ah-Counter, coach, or listener record filler patterns manually without turning on a microphone. Use it to identify when a habit recurs, then decide with context whether that pattern deserves attention.