The Mystery of Fillers.

Back in 2009 we had a lengthy discussion of “Filler words in different languages”; now Dan Nosowitz reports for Atlas Obscura on the linguistics of the subject:

Until about 20 years ago, few linguists paid filled pauses much attention. They were seen as not very interesting, a mere expulsion of sound to take up space while the speaker figures out what to say next. (In Russian, filled pauses are called “parasite sounds,” which is kind of rude.) But since then, interest in filled pauses has exploded. There are conferences about them. Researchers around the globe, in dozens of different languages, dedicate themselves to studying them. And yet they still remain poorly understood, especially as new forms of discourse begin popping up. […]

Though some researchers have insisted that filled pauses are individual words in their own right, with distinct meanings, many believe that there’s something more fundamental about them. With a few exceptions, filled pauses exist in every language, and are weirdly similar. In English, it’s “uh” or “um,” in Mandarin it’s “en,” in French it’s “euh,” in Hindi it’s “hoonm,” in Swedish “ohm.”

These are all very similar; essentially, they’re a centered vowel which may or may not be followed by a nasal consonant. […]

There are very few elements of language that are consistent amongst English, Mandarin, French, Hindi, and Swedish. And yet this one is pretty much the same.

We don’t really know where filled pauses came from, partly because, Twitter aside for the moment, they are oral sounds, and very unlikely to be found in historical written records. (Scholars have the same problem with swear words.) “Despite the lack of records about historical filler usage, it’s probably safe to assume that fillers have always been a part of human language,” says Katharine Hilton, a linguist at Stanford University who studies (among other things) filled pauses. “The reason for this is because they’re very useful words and communicate a lot of information to the listener.” The very earliest recordings of the human voice show that Thomas Edison was an avid user of “uh” and “um.” That’s about as far back as our data goes, but it seems fair to assume they go back further than that. These non-words, these mistakes, these errors: these are basic building blocks of language.

There’s interesting stuff about Japanese (where the most common fillers are ano and eto) and about second-language learning (Ralph Rose, a professor at Waseda University, “believes that filled pauses should be a significant part of language classes”). Thanks, Trevor!


  1. Most common filler sound in Russian is just plain ээээээээ, conviniently translated into English as aaaaaaaa. Uuuuuuh is also pretty common as well as a variety of open-mouthed sounds in between. They are not followed by any nasal coda. Contemplative mmmmmmm (not unkown in English as well) is there too, but is not preceeded by a vowel. In other words, if you are looking for some diversity, no need to go as far as Mandarinia.

  2. To be picky – I think “mid” may be a better descriptor than “central”, given that many of them involve [e] and [o].

  3. So. Western researcher, apparently unconcerned with (or unaware of?) the fact that SOV order is not unknown even among Indo-European languages, is pretty sure that this weird feature makes basic conversation in Japanese a cognitive burden. The science of n=1…

    (To be fair, this is reported second-hand, and perhaps Dr. Rose would have phrased things more carefully himself. Then again, he does praise the article on the home page of the referenced institute.)

    I sometimes feel like Eurocentricism and Japanese exceptionalism, despite what might seem to be an apparent mutual incompatibility, have basically settled into a tacit alliance when it comes to perpetuating such myths of Japanese Weirdness, which benefit both sides for entirely different reasons. You could literally consume a lifetime putting out the fires.

  4. Agree with Elessorn. People who study only two languages and feel qualified to make broad generalisations on that basis should try learning a third. It might put some of their pet theories to rest.

    A few other points.

    The Japanese is not eto, it’s eetto. I’m also trying to think whether Japanese people really use a long ii sound as a filler. Perhaps they do, but it seems to me that this is nothing more than a drawing out of the final vowel of the word (e.g. sonoooo…).

    Just as in English, where people use ‘you know’ and ‘like’, Japanese also uses other fillers, including nee and desu nee (e.g., Sore wa desu ne, … ‘It’s like this…’). I should think that there are lots of gradations between meaningless sounds and grammatical fillers, and it seems narrow to the point of parody to single out just a couple with reference to Japanese.

    Chinese has 那个 nèige ‘that’ (similar to ano). There are others like 就是 jiù shì or 就是说 jiù shì shuō ‘that is’, etc..

    Strangely enough, something like 那个 nèige also seems to be used in Mongolian. I’ve heard people pronounce it something the same (нэйгэ). It sounds for all the world like the Mongolians have borrowed it from Chinese, although I’m not sure how plausible that is. This filler can also be pronounced something like нөгөө meaning ‘the other one’, and that’s how it’s usually written (although writing it down isn’t so common).

  5. Elessorn: SOV order is extremely common worldwide, but Rel N – to which the text is clearly alluding – is a good deal rarer (see WALS). Long-distance dependencies, and in particular center-embedding, really do seem to be difficult to process cross-linguistically, and where they tend to come up depends on the language. There’s some very nice work by John Hawkins on the relevance of processing to syntactic typology, with specific discussion of Japanese, eg Processing typology and why psychologists
    need to know about it

  6. David Marjanović says:

    Then again, he does praise the article on the home page of the referenced institute.

    He does, or the university’s press department does?

  7. Heinrich R. Blutvergieben says:

    Mongols possibly borrowed [i]neige[/i]? Interesting.
    What are fillers like in sign languages? (I haven’t even read the article, perhaps it’s already answered there.)

  8. Sorry, how do you do italics here?

  9. In ASL I’m told one type of pause is to hold the ending position of a sign, sort of like the terminal Japanese vowel that Bathrobe describes. I believe this is different than a prosodic slowdown in utterance but I only know a little.

  10. Lameen

    Rel N – to which the text is clearly alluding – is a good deal rarer

    Perhaps, but you would need to demonstrate that all those Rel-N languages have the same type of filler behaviour as Japanese. Otherwise, the whole edifice falls down.

    As for the nature of ‘relative clauses’, there is some question that adnominal clauses in Japanese and relative clauses in English are fully comparable. It might not make a difference to the structural argument, but the looser way that Japanese clauses attach to nouns arguably makes them easier to manipulate and use. (See Grammar and Semantics of Adnominal Clauses in Japanese by Yoshiko Matsumoto.)

  11. My personal filler is /ˈθiːʌ/, presumably < the-uh, but I use it even where an article would be ungrammatical.

  12. In traditional RP English, the most common fillers were “I say, what,” and “as it were” and “as the bishop said to the actress.”

  13. Y.-R. Chao once related an anecdote of a Chinese principal going out of his depths trying to give a speech in German – that must be in the 30s or the 40s when German was still as known as English is today. Chao was puzzled by the frequent utterances of das in that speech, when it suddenly dawned upon him that the incomprehensible demonstrative is calqued from the Chinese nèige.

  14. It’s funny, I still have nèige and ano firmly implanted in my linguistic memory (from my time in Taiwan and Japan respectively), and occasionally produce them for my own entertainment.

  15. Athel Cornish-Bowden says:

    “as the bishop said to the actress.”

    I’ve always heard it the other way round: “as the actress said to the bishop”.

  16. January First-of-May says:

    In Russian, filled pauses are called “parasite sounds,” which is kind of rude.

    Pretty sure the phrase is “parasite words” (слова-паразиты).
    “Paraside sounds” would be “звуки-паразиты”, which I can’t recall having ever encountered (though maybe some do call them that).

  17. Pretty sure the phrase is “parasite words” (слова-паразиты).

    Yeah, that’s presumably one of those factoids people pick up and vaguely remember and then reproduce in garbled form.

  18. @tangent

  19. “звуки-паразиты”

    I was also surprised to see this expression, but a quick googling shows that that’s how they are called.

  20. I’ll be damned.

  21. > So then I went back…ahhh

    This example in the article is completely misleading, since it doesn’t explain that non-final Japanese prosodic units almost always end in vowels (since they usually end with particles, all of which end in vowels). So it’s really a repetition of the last sound, not the last vowel. But of course, stating it like that would take away from the weirdness that makes such a good story.

    > it seems to me that this is nothing more than a drawing out of the final vowel of the word (e.g. sonoooo…).

    This is a different phenomenon, I believe. An example is “ga” (stop), after which the speaker realizes they’re not ready to continue yet, and fills in with an “aaaaa”. It’s not a long ga.

    I associate this kind of fillers with speeches, which is where I usually hear them. I’d be suprised if people who use them also use them in their most colloquial speech.

Speak Your Mind