Filled Pause Choice as a Sociolinguistic Variable

Josef Fruehwald
University of Edinburgh

October 22, 2015
NWAV44 Toronto


Basic Points

  • Filled Pause selection (UM ~ UH) is a sociolinguistic variable, exhibiting internal and external conditioning.
  • There is a language change in progress towards selecting UM relative to UH.
  • The behavior of (UHM) appears to be exactly like that of any other sociolinguistic variable.
  • The paralinguistic nature of filled pauses is challenging to how knowledge of variation is sometimes construed in sociolinguistic research.

“Sociolinguistic Variable”

A “rule of grammar” (Labov, 1969).

  • Formal, Explicit (Labov, Cohen, Robins & Lewis, 1968) \[ \left\{\begin{array}{cc} \text{t}\\ \text{d} \end{array}\right\} \rightarrow \langle\emptyset\rangle/ \left[\begin{array}{rl} \alpha & \text{consonantal}\\ \zeta & \text{obstruent} \end{array}\right]~ \gamma(+) \delta(+)~ \begin{array}{c} \text{___}\\ \epsilon ~\text{voice} \end{array}~ \beta(\text{V}) \]


Defining the Variable

Variants Considered

  • UM: A neutral vowel, usually [ə], followed by labial nasal.
  • UH: A neutral vowel, usually [ə], without a following labial nasal.

Why Not Other “Fillers”?

  • Kendall (2013) found silences adjacent to (UHM) to be much longer than other fillers (well, like, y’know).
  • I’m following the lead of disfluency research in psycholinguistics that treat (UHM) differently from other fillers

Previous (UHM) Work

Psycholinguistic Perspective

Clark & Fox Tree (2002) summarize 3 perspectives:

  • filler as symptom: an “automatic, or involuntary” vocalization
  • filler as non-linguistic signal: fillers are part of a paralinguistic signaling stream.
  • filler as word: fillers are words like any other.

UM tends to be associated with greater processing difficulty (longer following silence).

Previous (UHM) Work

Sociolinguistic Perspective

Tottie (2011), Acton (2011), Laserna, et al (2014)

  • Women tend to have a higher UM:UH ratio.
  • Younger people tend to have a higher UM:UH ratio.

(UHM) Change

Wieling, et al (forthcoming) have found apparent and real time changes towards more UM use in

  • English
  • German
  • Dutch
  • Norwegian
  • Danish & Faroese.

(UHM) Change in the PNC

UH UM total
19,123 6,391 25,514

It is exactly what it looks like

A Generational Change

I fit a 2 dimensional non-linear mixed effects logistic regression model.

um ~ s(dob, year) + (1|speaker)

A Generational Change


  • Since UM is associated with greater processing difficulty maybe
    • people are signalling greater difficulties more often than they used to?
    • people are planning more complex utterances than they used to?
  • Maybe there is a new discourse function of UM, like turn initial UM?


Also D’Arcy (2012) on quoted thoughts and be like in NZE.

Eliminating some alternatives

As a rough measure of utterance complexity, proportion of function words out of all words.

Eliminating some alternatives

It looks like UM and UH are trading off in frequency

Eliminating some alternatives

The change seems to be happening in all discourse contexts


  • Since UM is associated with greater processing difficulty maybe
    • people are signalling greater difficulties more often than they used to?
    • people are planning more complex utterances than they used to?
  • Maybe there is a new discourse function of UM, like turn initial UM?

What is the change?

Usage Change

In any given context, both UM and UH are licit, but one may be preferred.

There is a shift in in the frequency with which one of two equally licit variants is selected.

Message Change

UM and UH are functionally differentiated to signal different messages.

There is a shift in the frequency of the message that is being communicated.

Silence Predictive Power

Under the changing messages hypothesis, the predictive power of the duration of following silences should be stable.

Under the changing usage hypothesis, its predictive power should decrease as speakers start using “um” more often irrespective of the context.

Silence Predictive Power

Filled Pause Choice as a Sociolinguistic Variable

It it quacks like a duck…

  • (UHM) exhibits a quantitative systematicity much like any sociolinguistic variable.
  • The basic rate at which a speaker uses UM is arbitrary, thus must be learned, thus constitutes an aspect of their knowledge of their language.
  • Accounting for (UHM)’s systematicity within a grammatical framework seems a bit far fetched.
  • \(\text{FilledPauseP}\leftrightarrow\text{ə}\langle\text{m}\rangle/\text{___}\alpha(\text{silence})\)

Not words like any other

(UHM) does not behave like other vowel initial words with respect to phonologically conditioned allomorphy.

a apple pear (UHM)
[ə] x 42
[ən] x 0
[ei] ? ? 112

“The Sociocultural Selector”

Sociolinguistic variables exhibit very similar quantitative systematicity across multiple levels of language:

  • Discourse
  • Syntax
  • Morphology
  • Phonology
  • Phonetics
  • Whatever inserts “UM”.

The nature of this systematicity may be best located outside of the formal properties of each of these systems, what Preston (2004) calls a “sociocultural selection device.”

(UHM) More Specifically

Colleagues in allied fields of psycholinguistics are very interested in filled pauses, the effect they have on listeners, and how they fit into models of speech planning.

The End


  • Acton, E. K. (2011). On gender differences in the distribution of um and uh. University of Pennsylvania Working Papers in Linguistics, 17(2), 2.
  • Clark, H. H., & Fox Tree, J. E. (2002). Using uh and um in spontaneous speaking. Cognition, 84, 73–111.
  • D’Arcy, A. (2012). The diachrony of quotation: Evidence from New Zealand English. Language Variation and Change, 24(03), 343–369. doi:10.1017/S0954394512000166
  • Kendall. (2013). Closer Looks at Speech Rate and Pause Variation. Palgrave Macmillian. doi:10.1057/9781137291448.0012
  • Labov, W. (1969). Contraction, Deletion, and Inherent Variability of the English Copula. Language, 45(4), 715–762.
  • Labov, W., Cohen, P., Robins, C., & Lewis, J. (1968). The Study of the Non-Standard English of Negro and Puerto Rican Speakers in New York City: Phonological and grammatical analysis: Report on Co-operative Research Project 3288. New York: Columbia University Press.
  • Laserna, C. M., Seih, Y. T., & Pennebaker, J. W. (2014). Um… who like says you know. Filler word use as a function of age, gender, and personality. Journal of Language and Social Psychology, 33(3), 328-338.
  • Preston, D. R. (2004). Three kinds of sociolinguistics: A psycholinguistic perspective. In C. Fought (Ed.), Sociolinguistic Variation: Critical Reflections (pp. 140–158). New York: Oxford University Press.
  • Tottie, G. (2014). On the use of uh and um in American English. Functions of Language, 21(1), 6–29. doi:10.1075/fol.21.1.02tot
  • Wieling, M., Grieve, J., Bouma, G., Fruehwald, J., & Coleman, J. (forthcoming). Variation and change in the use of hesitation markers in Germanic languages.