Filled Pause
Research Center

Filled Pause
Research Center

Filled Pause
Research Center

Investigating 'um' and 'uh' and other hesitation phenomena

Investigating 'um' and 'uh' and other hesitation phenomena

Investigating 'um' and 'uh' and other hesitation phenomena

January 31st, 2017

Great article in Atlas Obscura about filled pauses

Word cloud: filled pauses (produced using Wordle.net)

I was recently contacted by Dan Nosowitz, a writer for the web site Atlas Obscura, requesting an interview about filled pauses. I was pleased to accommodate and we arranged a Skype chat a few days ago. He had already written several other articles on linguistic topics (e.g., Canadian 'eh', Regional dialects), so it was an easy chat for me. We talked about many facets of filled pauses, sound, meaning, usage, acquisition, evolutionary origins, and more.

At the end of the conversation, I asked how soon he would publish an article and he said it would probably be a week later. To my surprise, about a day and a half later I noticed via Twitter that the article was published! If only I could turn out writing at that speed…

The article—titled "The Mystery and Occasional Poetry of, Uh, Filled Pauses"—is very well-written and engaging and I'm quite humbled to be featured in it. There is an image of science writers as being prone to making technical errors or exaggerating the claims of the researchers they interview. This was definitely NOT my experience in the present case. In fact, if there are any technical inaccuracies in the description of filled pause research, then these are due to a lack of clarity on my part during our chat. I'd like to address a couple of these as follows.

In our discussion of filled pauses in Japanese, I explained that there are some speakers who fill their pauses with a vowel sound matching the last vowel sound they uttered. To illustrate how this might look in English, Dan put the following example in his article (where the underlined vowels are the target vowels which the filled pauses matches).

*"So then I went back ... ahhh ... to my hometown ... oww ... to see my friends ... ehh ... who I haven’t seen in awhile."*

We briefly discussed this exact example on our chat, but I should have corrected it because it suggests that there could be an intervening consonant between the filled pause and the vowel it copies. Because of Japanese syllable structure, this never happens. Japanese syllables end in a vowel (tofu, sushi) or a nasal coda consonant (shinkansen, Fuji-san). Therefore, the copied vowel sound is actually the last sound uttered, a vowel (or, in a few cases, a nasal consonant (nn-). Thus, an English example that would more precisely reflect what's happening in Japanese would be the following.

*"John wants to know ... oh ... whether the fee ... ee ... could be paid through ... oo ... other means."*

Another point that I should clarify stems from the following portion.

There is a wide and contentious debate about what a filled pause even is. Ralph Rose … says he tends to use different definitions based on whatever he’s studying at that moment.

To the best of my memory, this is an accurate transcription of what I said in our chat. I think most researchers would understand what I mean, but I suppose others might take this to mean something like I change the definition of filled pauses to suit my needs (i.e., to get the conclusion I'm looking for). But what I meant is tied crucially to the first sentence. Because there are such diverse definitions of filled pauses across investigative fields, it is necessary to change one's operational definition in order to be comparable to an existing line of investigation. In acoustic phonetic studies, it's more common to limit filled pauses in English to only 'um' and 'uh'. But in sociolinguistic studies, one might wish to group them with discursive markers like 'like' and 'you know'.

Screenshot of definition of *interjection* from Wiktionary.com

Finally, the following portion just made me blush.

There is no evidence, anywhere, that the use of filled pauses is correlated in any way with any measure of intelligence. (Rose, in fact, describes himself as "a frequent ummer.")

So apparently, I’m to be taken as a counter-argument to the filled-pause-as-(non)intelligence-measure claim (how about we call this the "umintelligent hypothesis"). I doubt I’m particularly persuasive in this role, but I can appreciate the elliptical writing effect here as much as anyone.

Well played, Dan.